Can AI Build an Entire Business We Tried It Live

Can AI Build an Entire Business? We Tried It Live. 🔗 **Source:** [Can AI Build an Entire Business? We Tried It Live.](https://www.youtube.com/watch?v=wY2mXyuwu8I) > [!note]+ Document Notes > tags: #ai-development #web-development #business-automation #coffee-shop #loyalty-programs #no-code > author: ArrowHQ > source: YouTube > reference: https://www.youtube.com/watch?v=wY2mXyuwu8I > date created: 2024-12-28 11:30 > > # 🏷️ Title: Can AI Build an Entire Business? We Tried It Live. >  ## 🧰 Action Inbox - [H1] Test end-to-end loyalty program integration with database and email — 45:30 #business-automation - [H2] Compare AI builder costs vs traditional agency development — 52:15 #cost-analysis - [H3] Implement staff dashboard with QR code verification system — 32:45 #loyalty-programs ## 🧭 Table of Contents - [[#📝 Summary]] - [[#📖 Main Content]] - [[#🗒️ Notes]] - [[#📚 References]] - [[#🗣️ Paraphrase]] - [[#📰 Newsletter]] - [[#✅ Action Items]] - [[#🧪 Hypotheses]] - [[#📊 Methodology]] - [[#❓ Questions]] - [[#🔮 Future Research]] - [[#💡 Implications]] - [[#🧠 Learning Styles Assessment]] - [[#📊 Accessibility Assessment]] - [[#🌍 DEIB Assessment]] - [[#✅ Logical Validity]] - [[#🛡️ DARVO Analysis]] - [[#🎯 Conclusion]] ### 🔗 References - [Arrow AI Builder](https://arrow.ai) - [Getty Images Integration](URL) ## 📝 Summary - Two developers simultaneously built complete coffee shop websites using Arrow AI Builder in 15 minutes - Both started with identical prompts but created vastly different solutions with unique features - The AI automatically generated complex loyalty programs, ordering systems, and databases without technical expertise - Arrow AI introduced error handling that doesn't charge credits when the AI causes mistakes - The demo revealed how AI can create full business plans including marketing strategies and growth frameworks ## 📖 Main Content This live demonstration showcases Arrow AI Builder's capability to generate complete business applications from simple prompts. Two developers, starting with the identical brief of building a coffee shop website with online ordering and loyalty programs, ended up with completely different implementations - one in Dubai with an Australian theme, another in New York with a "hippie and stylish" aesthetic. The AI builder automatically generated sophisticated features including multi-tier loyalty programs with creative naming (Seedling, Bloom, Rooted), automated email confirmations, staff dashboards, and integrated databases. Beyond just building websites, the tool created comprehensive business strategies with marketing plans, customer retention frameworks, and operational workflows. A key product update was revealed during the build: Arrow AI now detects when errors are caused by the AI agent versus user input, automatically fixing agent-generated errors without charging credits to maintain fairness for customers. ## 🗒️ Notes #### ⏱️ [00:00](https://youtu.be/wY2mXyuwu8I?t=0) — [H1] Starting the simultaneous build challenge - Key idea: Both developers begin with identical coffee shop prompt but expect different outcomes - Evidence: "Same started, same starting point, but completely different results" #### ⏱️ [02:15](https://youtu.be/wY2mXyuwu8I?t=135) — [H2] AI automatically generates business plans and naming - Key idea: Arrow AI creates comprehensive project plans including branding decisions - Evidence: "It went ahead and did it by itself. So, it's like homepage, full menu page, online ordering integration, loyalty program page" #### ⏱️ [05:30](https://youtu.be/wY2mXyuwu8I?t=330) — [H3] Localization and currency adaptation - Key idea: AI automatically adapts to geographic context in prompts - Evidence: "It's taken local currency, which I love. So, because I've said it's based in the UAE and Dubai" #### ⏱️ [08:45](https://youtu.be/wY2mXyuwu8I?t=525) — [H1] Error handling and credit fairness system - Key idea: New feature prevents charging customers for AI-generated errors - Evidence: "We will assume the cost of the credits there versus charging the customer for the credits" #### ⏱️ [12:30](https://youtu.be/wY2mXyuwu8I?t=750) — [H2] Direct editing capabilities preserve credits - Key idea: Manual text editing bypasses AI processing to save computational costs - Evidence: "You can go in and directly edit the text... you don't consume credits which is great" #### ⏱️ [18:20](https://youtu.be/wY2mXyuwu8I?t=1100) — [H1] Complex loyalty program generation - Key idea: AI creates sophisticated multi-tier loyalty systems with creative branding - Evidence: "Seedling, bloom and rooted... It's giving you marketing ideas within the build" #### ⏱️ [25:45](https://youtu.be/wY2mXyuwu8I?t=1545) — [H2] Database integration and technical architecture - Key idea: AI explains backend requirements and guides technical implementation - Evidence: "To make it truly functional, you need a backend API route to receive and store signups" #### ⏱️ [32:15](https://youtu.be/wY2mXyuwu8I?t=1935) — [H3] Staff dashboard and verification systems - Key idea: AI designs practical in-store verification workflows - Evidence: "Staff scans a QR code and adds a stamp. Simple, no customer account needed" #### ⏱️ [38:50](https://youtu.be/wY2mXyuwu8I?t=2330) — [H2] Email integration and business identity requirements - Key idea: Professional email setup becomes crucial for customer trust - Evidence: "You'd really need to have your info at or hello at Soho Brew" #### ⏱️ [45:30](https://youtu.be/wY2mXyuwu8I?t=2730) — [H1] End-to-end functionality testing reveals gaps - Key idea: AI-built interfaces require thorough testing to ensure full functionality - Evidence: "You have to go in and test it. you have to build it beyond just the mockup" #### ⏱️ [48:15](https://youtu.be/wY2mXyuwu8I?t=2895) — [H3] Domain registration and business launch capabilities - Key idea: Platform integrates domain purchasing for complete business setup - Evidence: "Hey there, soob brws.com is available... You can buy your domain and everything in one place" #### ⏱️ [52:15](https://youtu.be/wY2mXyuwu8I?t=3135) — [H1] Cost comparison with traditional development - Key idea: AI builder delivers agency-level results at fraction of traditional cost - Evidence: "Building a full on loyalty program that you would usually pay hundreds or thousands of dollars" ## 📚 References - [Arrow AI Builder](https://arrow.ai) - [Getty Images Integration](https://getty.com) - [SMTP Email Setup Documentation](URL) ## 🗣️ Paraphrase - AI builders can generate complete business applications including sophisticated loyalty programs and databases from simple prompts in minutes rather than weeks - The same starting prompt can produce vastly different solutions, demonstrating AI's creative interpretation capabilities and the importance of iterative refinement - Professional business deployment requires additional considerations like domain registration, email setup, and thorough functionality testing beyond the initial AI build ## 📰 Newsletter **AI Builders Are Revolutionizing Small Business Development** This week, we witnessed something remarkable: two developers built complete coffee shop businesses with ordering systems and loyalty programs in just 15 minutes using AI. What started as identical prompts became unique solutions - one Dubai-based with Australian flair, another New York-focused with creative branding. The standout feature? Arrow AI's new error handling system that detects when mistakes are the AI's fault, not yours, and fixes them without charging credits. Plus, the integrated database creation, email systems, and even domain purchasing mean you can go from idea to live business faster than ever. The practical takeaway: AI isn't just building websites anymore - it's creating comprehensive business strategies, marketing plans, and operational workflows that would typically cost thousands from agencies. Resources: [Arrow AI Builder](https://arrow.ai), [No-Code Development Guide](https://learn.wordpress.org/), [Small Business Automation](https://developer.wordpress.org/) ## ✅ Action Items #### DevRel - Developer advocacy: Create tutorials showing AI builder integration with existing WordPress workflows - Education: Develop workshop curricula for no-code business development using AI builders - Developer experience: Document API integration patterns for loyalty systems and databases - Community management: Facilitate discussions on AI development ethics and credit fairness #### OSPO - Policy & Governance: Establish guidelines for AI-generated code licensing and attribution - Software Inventory: Track AI builder dependencies and generated code provenance - Developer Support: Provide training on AI-assisted development best practices #### Community - Maintainers: Support integration between AI builders and WordPress core features - Contributors: Create plugins that bridge AI-generated sites with WordPress ecosystem - Ecosystem: Coordinate standards for AI-generated business applications #### WordPress - Core: Evaluate AI builder integration opportunities for WordPress.com and .org - Marketing: Develop narratives around WordPress as platform for AI-enhanced businesses - Documentation: Create guides for integrating AI-built applications with WordPress backends - Hosting: Establish performance guidelines for AI-generated high-traffic applications #### The WPCC - Programs: Fund development of open-source alternatives to proprietary AI builders - Projects: Support documentation of AI development best practices for small businesses ## 🧪 Hypotheses - AI builders can reduce small business website development costs by 80-90% compared to traditional agencies - Identical AI prompts will produce functionally similar but aesthetically diverse results across multiple runs - Non-technical users can successfully deploy complex business applications using conversational AI interfaces - AI-generated business applications require human oversight for security and regulatory compliance ### 📊 Methodology - Comparative cost analysis between AI builder outputs and equivalent agency quotes - A/B testing of identical prompts across multiple AI platforms and sessions - User experience studies with non-technical small business owners using AI builders - Security audit framework for AI-generated business applications ### ❓ Questions - How do AI builders handle industry-specific regulations like PCI compliance for payment processing? - What happens to AI-generated businesses when the underlying platform changes pricing or features? - Can AI builders scale to enterprise-level complexity while maintaining the simplicity demonstrated? ### 🔮 Future Research - Integration patterns between AI builders and established CMS platforms like WordPress - Governance frameworks for AI-generated business applications in regulated industries - Long-term maintenance and evolution strategies for AI-built applications ### 💡 Implications - Practice: Small businesses can now deploy sophisticated digital infrastructure without technical teams - Policy: Regulatory frameworks need updating for AI-generated business applications and data handling - Security: AI-built applications require new audit approaches for generated code and database schemas - Education: Business education must include AI-assisted development as core entrepreneurship skill - Funding: Venture capital and small business lending models may shift toward AI-enabled rapid prototyping ### 🧠 Learning Styles Assessment Reading level estimate: 8th Grade (Flesch-Kincaid) #### 👀 Visual learners - Present: Screen recordings of live building process, UI mockups, color scheme demonstrations - Improve: Architecture diagrams showing database relationships, workflow visualizations #### 🔊 Aural learners - Present: Live commentary and discussion during building process - Improve: Podcast-style summary explaining technical concepts and business implications #### 🗣️ Verbal learners - Present: Conversational AI interaction examples, terminology explanations - Improve: Glossary of technical terms, business concept analogies #### 🧩 Kinesthetic learners - Present: Step-by-step building demonstration, hands-on prompt crafting - Improve: Interactive tutorial checklist, sandbox environment for testing #### 🧠 Logical learners - Present: Technical architecture explanations, cost-benefit analysis - Improve: Decision tree for choosing AI builder features, performance metrics dashboard #### 🧑‍🤝‍🧑 Social learners - Present: Collaborative building session, peer comparison of results - Improve: Community showcase of AI-built businesses, mentorship matching #### 🧘‍♀️ Solitary learners - Present: Individual building challenges, self-directed exploration - Improve: Personal project reflection prompts, progress tracking worksheets #### 🌿 Naturalistic learners - Present: Business ecosystem thinking, organic growth strategies - Improve: Sustainability impact analysis, local business integration patterns ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Clear heading structure in AI builder interface, descriptive button labels - Improve: Alt text for generated images, ARIA labels for dynamic content #### Keyboard navigation - Present: Tab navigation through builder interface - Improve: Skip links for complex generated forms, focus management in dynamic content #### Contrast and readability - Present: High contrast in builder interface - Improve: Ensure AI-generated sites meet WCAG contrast requirements by default #### Text resizing and reflow - Present: Responsive layouts in generated sites - Improve: Test generated content at 200% zoom, prevent horizontal scrolling #### Media and captions - Present: Video demonstration with spoken audio - Improve: Auto-generated captions for technical explanations, audio descriptions #### Forms and interactions - Present: Clear form labels in loyalty program signup - Improve: Error handling accessibility in AI-generated forms, validation announcements ### 🌍 DEIB Assessment #### Cultural breadth - Present: Multiple geographic contexts (Dubai, New York), cultural business themes (Australian coffee culture) - Improve: Examples from diverse global business contexts, multi-language AI builder support #### Inclusive language - Present: Gender-neutral business terminology, accessible technical explanations - Improve: Avoid assumptions about technical background, provide multiple complexity levels #### Neurodiversity support - Present: Clear step-by-step process, visual demonstrations - Improve: Adjustable pacing controls, alternative input methods for AI prompts #### Representation - Present: Diverse geographic and cultural business examples - Improve: Include accessibility-focused business examples, underrepresented entrepreneur showcases #### Psychological safety - Present: Supportive tone around technical challenges, error normalization - Improve: Explicit encouragement for non-technical users, failure-positive learning framing ### ✅ Logical Validity #### Strong reasoning - Deductive strength: Clear cause-effect relationships between AI prompts and outputs demonstrated throughout - Inductive strength: Solid evidence from live demonstration supports broader claims about AI capabilities - Causal reasoning: Well-established connections between feature complexity and development time/cost - Analogical strength: Effective comparisons between AI development and traditional agency work #### Fallacies - Appeal to novelty: Potential overemphasis on AI being "new" rather than focusing on practical utility - Hasty generalization: Limited sample size (two builds) used to make broad claims about AI capabilities #### Tone of voice - Overall tone: Enthusiastic and exploratory, with genuine surprise at capabilities and honest assessment of limitations ### 🛡️ DARVO Analysis - Deny: No denial patterns present - Attack: No attacking behavior observed - Reverse victim/offender: No victim/offender dynamics - Emotional cues and rhetoric: Positive excitement about technology capabilities, honest acknowledgment of complexity ## 🎯 Conclusion AI builders like Arrow AI are democratizing business application development by enabling non-technical users to create sophisticated systems in minutes rather than months. The live demonstration proved that identical starting prompts can produce diverse, functional solutions including complex loyalty programs, databases, and payment systems. However, success requires understanding that AI generates starting points that need testing, integration, and business-specific customization. The most significant barrier to small business digitization may no longer be technical complexity or cost, but rather knowing what questions to ask and how to iterate effectively with AI tools. ## 🧱 Refactor Appendix ### 📚 Literature Split Pad Purpose: Stage CEE units that become Literature notes. Split level: H3 ### [H1] AI builders reduce technical barriers for small business development > [!quote] [Can AI Build an Entire Business? We Tried It Live.](https://www.youtube.com/watch?v=wY2mXyuwu8I) by ArrowHQ via YouTube. Published 2024. Accessed 2024-12-28. > "People who wouldn't I mean I wouldn't have been able to do this... it would have taken time and reading and understanding." Demonstrates how AI eliminates traditional technical prerequisites for business application development. [view highlight](https://youtu.be/wY2mXyuwu --- ## Full Transcript [00:00] I read one hook, you read another, and [00:02] it's okay. [00:03] >> I did the same thing, though. So, you [00:04] want you want to do it like you want to [00:06] read the book first. [00:07] >> Yeah. We picked one business idea. We're [00:10] both building it live at the same time, [00:12] talking through every prompt as we go. [00:14] Same started, same starting point, but [00:17] completely different results. Let's do [00:19] it. [00:19] >> Okay, let's go. Let's see where we end [00:22] up. My suggestion is we build a [00:24] neighborhood co coffee shop. um has an [00:27] online ordering with online ordering and [00:30] a loyalty program. Let's start. I have [00:33] AeroBuilder open. So, I'm going to start [00:35] by sharing my screen. Yours is in [00:37] Arabic. [00:38] >> Yeah. [00:39] >> Yeah. Because I'm going to be testing [00:40] Arabic today. It's going live. [00:42] >> In Arabic. Mine came. Is this the You [00:44] just need to click on it in the [00:45] navigation. All right. We're back. So, [00:47] we said So, we said a website for a [00:50] local coffee shop. [00:52] >> Yeah. with a online ordering system and [00:55] a loyalty program. Do we're not are we [00:57] like going rogue on where it's where [00:59] it's located? [01:01] >> Yeah. Okay. I have a lowkey coffee shop [01:03] in the Lakes Dubai. Is there a We know [01:07] our coffee. [01:08] >> I'm building a local coffee shop in New [01:09] York. [01:10] >> Wow. [01:10] >> That's my That's like my my dream [01:12] location for a coffee shop. Who knows? I [01:15] think it's going to come up with like a [01:16] very local authentic [01:18] >> I mean, we need one here. Maybe it'll [01:19] take me to a new adventure again. [01:22] Exactly. [01:23] >> The start of a business. [01:24] >> Yeah, it could be. We're an Australian [01:26] that word he [01:26] >> loyalty program. There's a loyalty [01:29] program. [01:29] >> No, I'm going to make a [01:30] >> dessert. She's gone straight to [01:32] desserts. [01:32] >> You've gone past the coffee. [01:34] >> Of course, it's the one thing that I [01:35] know how to do the best. Okay, I think [01:37] I'm ready. That's I'm going to give us [01:38] like 15 minutes to build and then I'm [01:41] done. Yeah, I've really gone into the [01:43] detail of the coffee. No, you start at [01:45] taking time with my prom. [01:47] >> It's 1:33 now. [01:48] >> Okay. [01:49] >> What's the coffee shop called? So, Lena, [01:51] are you still writing your [01:54] >> I'm so proud of you. So, mine asked me a [01:56] few questions and now it created like a [01:58] plan for me. [01:59] >> Okay. Oh, did you ask it to create a [02:01] plan or [02:02] >> No, no, no. It it went ahead and did it [02:04] by itself. So, it's like homepage, full [02:06] menu page, online ordering integration, [02:09] loyalty program page, about and a story [02:12] page. [02:13] >> Mine's still thinking. It's named at the [02:15] Lakes Coffee. [02:16] >> The Lakes Coffee. [02:17] >> Yeah. So, it came up with a name. It [02:19] did, which I would might change later, [02:21] but that's a small detail cuz I haven't [02:22] committed to a name yet. [02:24] >> Okay, we can chat. [02:26] >> I I came up with the name Soho Brew. [02:28] >> Okay. I wonder if it's exists elsewhere. [02:31] These are the things where [02:34] I might run into a real Soho Brew coming [02:37] after me. [02:38] >> You may. [02:39] >> You'll realize when you try to register [02:40] your domain name, is it taken or isn't [02:42] it taken? That is the question. [02:44] >> You know, it's going to tell me right [02:45] now. And like AI builder, it has this [02:48] like in the domain URL, it recommends a [02:50] domain name that is available for me to [02:52] buy. [02:52] >> Oh, I like that. Let's try. [02:54] >> So even when I ran into my error, what I [02:56] said is like try again and it's just [02:59] like build rethinking. [03:01] >> Yeah. So I it it gave me a little plan [03:03] and now it's serving me with coffee with [03:05] images actually for the homepage build. [03:08] It's given me a color scheme and font, [03:11] but still thinking I don't like the [03:12] colors that it's given me that much. [03:14] Like I like a more pastel. [03:16] >> Which colors did it give you? [03:17] >> Like a brown color palette. Um, so [03:20] >> mine is like brown with greens. [03:22] >> I don't have It's done brown, black, and [03:24] like a rusty type color, which I'm not. [03:27] >> Which font did it give you? [03:29] >> Uh, Comarant Garamond plus Leato. [03:32] >> Me too. [03:33] >> And I think this is where I find you [03:36] probably have to add the specifics in [03:38] the design to get what you want if you [03:41] like a certain font. [03:42] >> Yeah. cuz like I know what I don't want [03:44] as well. So like this color scheme I [03:46] said make it like modern and this huga [03:48] feel you know that Swedish comfort, but [03:51] I think it's gone to um it's [03:54] overorrected to index on the copy colors [03:57] versus something more pastel. [03:58] >> I asked for like a hippie and stylish. [04:01] >> Oh, I'm curious to see what that comes [04:03] up. [04:03] >> Yeah, me too. [04:05] >> And you know what I realized? You know [04:06] when you and I both did the [04:07] International Women's Day? I'm going off [04:09] topic here, but we did the International [04:10] Women's Day. You gave us both a black [04:12] and pink. Yes. No, mine was green and I [04:16] asked her to change it to pink. I think [04:18] I did the same. Is it weird that we [04:20] defaulted it? Like I think it gave me [04:22] the option of green and pink and I [04:24] wanted pink. [04:26] >> Okay. Mine actually built something for [04:28] me and I really really like it. [04:30] >> Really? [04:31] >> Yeah. [04:32] >> Oh, good for you. I'm happy. Oh, wow. [04:34] It's just up. So, it gave me a little [04:37] preview that I didn't like that much. [04:39] And then I've went back and thought [04:40] about it. And I'm not sure if I hate the [04:42] color scheme as much as I originally [04:44] thought I did. [04:45] >> I'm not as passionate about the font. [04:48] >> I've got chat with Arrow. It's saying, [04:49] "Tell Arrow what you want to build or [04:51] modify in your app. The AI will [04:52] understand your requests and make [04:53] changes for Okay, thank you. [04:55] >> Mine is saying something else." [04:56] >> What is it saying? [04:57] >> What do you think of the design? [04:58] >> It's more confident. Look at you. [05:00] >> Yeah. Do you think it's the end the [05:02] prompt that we end with? [05:03] >> I don't know. Something that was [05:05] recently launched and now I'm seeing it [05:07] and I'm so happy about it was that there [05:09] was a runtime error [05:11] >> and then the chat prompted itself to fix [05:14] and analyze the runtime error and then [05:17] it says completed for free. [05:19] >> So it corrected the error and I was not [05:22] credits were not charged and this is I [05:24] think it's important to pause on this [05:26] one where you know technology will have [05:28] errors. It's like when you have a bug on [05:30] your phone and you have to restart or [05:32] you have to restart your laptop. What [05:33] we're really trying to do is recognize [05:35] when Arrow AI builder has an error in [05:38] credit or is the cause of the error. So [05:40] it's not a userenerated error. It the [05:42] tool it's an agent generated error. And [05:45] so we will assume the cost of the [05:46] credits there versus charging the [05:49] customer for the credits. And so that's [05:51] now implemented in. So it's just a check [05:54] mechanism to ensure that we continue to [05:56] allow like fairness and build for the [05:58] customer. So, we'll own the [06:00] responsibility for uh for generated by [06:04] the agent and let you to continue to [06:06] build as desired with the prompts, but [06:08] it won't it won't fix your you know [06:10] quality prompts. You still have to go [06:12] and fix that yourself. [06:13] >> Yeah. And like kudos to the product [06:14] manager for sure [06:16] >> that came up with this idea cuz we we [06:18] have seen a lot of feedback from [06:20] consumption of credits and it was a [06:22] really thought through process of like [06:25] how can we fix this? I mean it's it's a [06:27] true example of putting the customer. I [06:29] think it's two things. It's one going [06:31] deeply into the product and [06:32] understanding where the product is not [06:35] meeting the bar in terms of there are [06:36] certain tasks that need to be iterated. [06:38] And again like technology develops on a [06:41] constant basis you know especially in [06:43] this agentic age but at the same time to [06:46] the product manager who said hey this is [06:48] an opportunity for us to do what's right [06:50] for the customer. It was a very it was a [06:53] very conscious process to say okay let's [06:55] own accountability for what our tool [06:57] does what our agent does and um and make [07:00] sure that the customer isn't penalized [07:01] for it. So it's a proud moment like I [07:03] feel I feel very proud to be on a team. [07:06] >> Yeah. Thanks Josh out. [07:07] >> I will make some changes to my website [07:09] now. [07:10] >> So it's asked me so it's actually for me [07:12] it's gone and said what do you think of [07:13] the design? And rather than going and [07:15] saying I don't like it I'm first going [07:16] to play with the color scheme and then [07:18] I'm going to adjust the [07:19] >> oh I'm going to do the same thing. I'm [07:21] going to change my browns to burgundy. [07:23] >> Did you get Oh, maybe you would have [07:24] liked my design better. [07:25] >> Does yours have like burgundy in it? [07:27] >> It's got this like rusty orange. Do you [07:29] think it's it's triggered by the copy? [07:31] >> I think it's triggered by the vertical [07:33] that we're building for. [07:34] >> Me, too. [07:35] >> And probably it goes and finds like best [07:37] colors that fit this vertical. So, I've [07:39] I've gone in and I've said I don't like [07:41] the colors. I want turquoise and greens [07:42] across the page. Basel tones. Mine is [07:45] like I said change the brown to [07:47] burgundy, but what I meant to say was [07:49] change the [07:51] >> burgundy. [07:52] >> Can you You have to undo it now. [07:54] >> I can undo it. I can actually go back to [07:56] >> the previous [07:57] >> to the backup and change it, but I like [08:00] what I came up with, so I'm not hating [08:02] it. [08:02] >> Sometimes the right road is the wrong [08:05] road. [08:06] >> Yeah. Oh, I love it. Now it's like Mine [08:07] is like burgundy, green, and nude. [08:10] >> Interesting. Yeah, we're still I'll [08:13] rewrap the page with the new turquoise [08:15] and cream palette brow and update the [08:16] header. And please do I talk a lot to [08:20] the agent when I build. Yeah, you know, [08:22] I feel like it needs to be more [08:24] personable. I usually say like do this, [08:27] do that. Sometimes when it's like not [08:28] doing what I want, I'd say please. I [08:30] type in please. I know it's a um it's a [08:33] no no because it uses credits and [08:35] compute and so in the grand scheme of [08:37] environment it's not a great thing but I [08:40] still feel like I don't want it to be [08:42] death to politeness of humankind. So I [08:45] feel like I still need to you know [08:47] invest in the etiquette that we've been [08:49] taught but don't do what I do because [08:51] you will eat your credit. [08:52] >> I just changed my font so I I asked it [08:56] asked it to change the font. It gave me [08:58] some options. I did change it. Let me [08:59] look at the new one. See? So, I updated. [09:02] I don't think it's fully done. I think [09:03] this color scheme does not work, but I'm [09:05] not going to overinvest time in this [09:07] because we're already 10 minutes into [09:09] the build and we said 15 minutes and I [09:12] haven't been able to do I don't have the [09:14] online. I think I'm going to commit [09:15] direct. I'm going to skip the font. I'm [09:17] going to skip the color scheme. I'm [09:19] going to go straight to the build of the [09:21] other pages so we actually see what it [09:22] does. [09:23] >> Okay. I'm I'm also building the other [09:24] pages, but I'm I'm quite impressed by [09:27] the speed. It's actually faster than [09:29] what I expected. [09:31] >> But I think it's important to point out [09:32] here the speed when you look at the [09:35] speed of different builders. I think [09:37] where there's a quite a large difference [09:39] is the detail at which it's trying to [09:41] make the change. So mine's just [09:43] reloading now and it's actually redone [09:46] the whole page in the new color palette. [09:48] >> Um, so it took a little bit of time but [09:50] the app but I'm just going to accept it. [09:52] I don't think I made the right color [09:53] choice but I love this section here. [09:55] It's super clean. change the font cuz I [09:57] just don't like this font. I wouldn't [09:58] choose this font. And not because I [10:00] don't think it looks nice here, but I [10:02] always go to physical print. So, let's [10:04] say this is a real physical shop. This [10:06] kind of font, I'm always a little bit [10:08] iffy on if I have to print anything, [10:10] >> it may not show up. I like more of the [10:13] the clean and modern fonts just because, [10:16] you know, whether you team it up with a [10:17] logo or text, they're generally a little [10:19] bit more flexible in in output. Don't [10:22] like these pictures. So, I would go back [10:24] and change these into much more gourmet [10:26] looking fresh produce. But I like this [10:29] little picture down here. Every cup [10:31] counts. Oh, this is so So, even on the [10:33] homepage, can do I'm curious to see what [10:36] your loyalty looks like. But it's got a [10:38] little loyalty program. Every cup [10:40] counts. So, earn a stamp with every [10:41] coffee. Hit 10 and get your next one. No [10:44] app required. Just your a good habit. [10:46] I'd asked it to do a point system, but [10:49] it's giving me a stamp system. So, it's [10:51] default. So, there's a little bit of a [10:53] disconnect, but it's taken my prompt to [10:56] the 25, 50, and 100 cups cuz I'd said [11:00] you hit like you get rewards at specific [11:02] milestones. So, then it said you get [11:04] bonus on a on top of the additional cup [11:07] and free coffee, you get more on top of [11:10] it. So, it's already giving me new ideas [11:12] in terms of what I could do. But, I'm [11:15] going to Oh, but it must tweak the shade [11:18] of the turquoise. You heard me, didn't [11:19] you? You heard me. Maybe I'll I'll have [11:22] to call it something. [11:23] >> Okay. So, right now I did build [11:26] >> See yours. Let me have a look at yours. [11:28] >> I did build the menu page. So, let me [11:32] show you. This is what my This is what I [11:34] the [11:34] >> I love the colors. [11:36] >> So beautiful. [11:37] >> It's so nice. I'm slightly jealous. I [11:39] feel like because you put matcha in the [11:40] mix, it took the Look, we've got the [11:42] same picture though here. [11:43] >> The the this one [11:45] >> layout. Yeah. [11:46] >> The loyalty program. I gave it No, I [11:49] love the burgundy. It looks lovely. [11:51] >> And then the menu is actually really [11:53] clean. [11:54] >> And when when we were talking about [11:56] changing the text without um without [11:59] without using credits, basically this is [12:01] it. When you see this like little star [12:04] over here, you click on it [12:05] >> and then you can go and edit directly. [12:07] >> I can remove this endlessly. [12:09] >> And so this is where that's actually [12:10] really cool. So people who are used to [12:12] building in a template format, they can [12:15] go in and directly edit the text. Yeah. [12:17] So if you have you know an existing [12:19] website or anything that you have [12:21] prepared going back to the discussion [12:23] previously that we had around planning [12:24] the content you can actually take it and [12:27] copy and paste it here cuz in this [12:29] scenario I would assume that you know [12:31] what's on your menu. So you can go in [12:34] and just directly explain what that [12:36] drink or what that beverage is to make [12:38] sure that it truly you know the coffee [12:40] type the beans etc. And you don't [12:44] consume credits which is great. So you [12:45] can be very smart about how you optimize [12:47] your usage, but also you can make the [12:50] build as simple or as complex as you [12:52] want to based on what you're comfortable [12:54] with. Yeah. And I think what the other [12:56] thing is that like changing images can [12:59] be so tricky with the agent [13:02] >> cuz I tried it sometimes like drives me [13:04] crazy. When you hover over the picture, [13:07] click on replace, you can either replace [13:09] it with one of the Getty images that we [13:11] have or you upload an image and change [13:13] it with it. It's so easy. It's really [13:15] simple. So, you don't have to. And that [13:17] saves you from having to write in a a [13:20] really exact prompt to say, "This is [13:22] exactly where I want you to take this [13:24] picture and put it in this placement on [13:26] the page. You can you can get to your [13:29] target position and your target image [13:31] and put it in there." [13:32] >> Yeah, [13:32] >> I think that's great. And it's it's also [13:34] good. You don't actually have to have [13:35] the images. You can go in and use Getty, [13:37] which makes it you don't need to have [13:39] anything prepared. You just need to know [13:41] how to utilize it. So it's done [13:45] order ahead pick up ready. So it's doing [13:48] the ordering. It's quite detailed. I [13:50] think we even though seems very simple [13:53] actually quite a detailed task for the [13:56] um for the agent to fill out. So you can [13:58] see we've gone to on this one we're [14:01] building out the um Oops. I've got an [14:03] error now. So it's fixing it. Please [14:05] analyze and fix this runtime error. I'm [14:07] going to do like a big change now. I'm [14:09] changing my header image to be a video. [14:12] Wow. Cuz I enjoy No, no, I enjoy [14:15] websites that have like [14:18] >> interacting. [14:18] >> Yeah, [14:19] >> you're such a next generation. I'm like, [14:21] don't overwhelm me with it. [14:23] >> So, while you do that, I'm going to go [14:25] and check out my ordering system, which [14:27] is built out. [14:28] >> So, here I actually love this. So, I [14:30] have an ordering system. It's taken my [14:33] queue on cortados and piccolo. I have [14:36] everything except a double espresso. [14:38] But, let's have a piccolo. I would I [14:41] could go back and [14:42] >> remove the matcha. [14:43] >> Remove the matcha. We don't serve [14:44] matcha. And then it's given me a little [14:46] option on the baked goods. But this is I [14:49] think again it goes back to the detail [14:50] of when you have your menu. You can then [14:52] input exactly what it that you offer. [14:54] Bliss ball and sea bar. Okay. I could [14:56] really go for a bliss bowl right now. So [14:58] here piccolo and then I can say Bieber [15:01] as soon as possible or I can say I would [15:03] like to go for a coffee in 30 minutes. [15:06] And the notes I'm going to say coconut [15:08] milk side of sugar. And then let's see [15:11] what happens when I try place order. [15:13] Thanks Selena Bieber. The order is [15:15] confirmed. And so that is and then [15:18] payment is at the counter. So what we [15:20] could do is we could go an extra mile [15:22] and look at if I can do online payments [15:24] so that the coffee shop doesn't prepare [15:26] the order and not get payment [15:28] fulfillment. I think it's always safer [15:30] to take the cash first. But that would [15:32] be the next that would be another phase [15:34] after I build the initial blueprint. But [15:36] if I go back to home, then it takes me [15:38] back to the homepage. So I'm going to [15:40] look at the loyalty program now and see [15:43] what it gives me. I'm going to let you [15:46] screen share. Are you screen sharing [15:47] now? [15:47] >> I can start screen sharing. I'm so [15:50] unhappy because my video is not working. [15:52] >> Okay. So what I'm doing now is checking [15:55] my loyalty program to see if it actually [15:58] works. And I said, "Can you check if the [15:59] loyalty program actually works?" And [16:01] then it says, "To make it truly [16:03] functional, you need a backend API route [16:06] to receive and store signups. You need a [16:08] database and a confirmation email sent [16:10] on signups." So I'm like, [16:12] >> that's a lot. He doesn't know how to do [16:14] it. [16:15] >> But it actually asks you how do you want [16:17] to proceed and it has the has it planned [16:19] out for you. So you can start by I want [16:21] to start by creating a database. And [16:24] what is really really cool is that with [16:26] the database, it's over here. Like it's [16:29] all inside the builder. You don't have [16:31] to know how to create an a database or [16:33] you don't have to have it externally. I [16:35] can show you now while it's being built. [16:37] But what you're effectively saying is [16:39] the technical barrier doesn't exist [16:40] anymore. Yeah. So people who wouldn't I [16:43] mean I wouldn't have been able to do [16:44] this. Maybe I would let me not [16:45] underestimate myself. But it would have [16:47] taken time and reading and understanding [16:50] to be able to get the database, [16:52] integrate it, you know, capture leads [16:54] and then integrate it into the site as [16:56] an order, a full end to-end loyalty and [16:58] ordering system. And then you want to [17:00] have the engagement. You want to be able [17:01] to contact those customers, you know, [17:03] share them, share information and [17:05] promotions with them. But you, the [17:08] agent, the builder does it for you now. [17:09] You just have to give it the right [17:10] instructions. So that's that's [17:12] incredible. [17:12] >> Yeah. And when you think about it, you [17:14] think, "Oh, I want to start a loyalty [17:16] program." But did you really think it [17:17] through of what it takes to have a [17:19] loyalty program? All the steps that you [17:20] need to have in your website to be able [17:22] to get the loyalty program to work. And [17:24] I think I like that it's telling me to [17:27] have a loyalty program that is actually [17:29] working. You need to do this and that [17:30] and this [17:31] >> and let's do it together. Let's build it [17:33] together. [17:33] >> Well, it allows you to understand the [17:35] steps or the processes required to make [17:37] a habit. And this isn't just for a [17:39] coffee shop. I mean, this could be a [17:41] community. This could be a, you know, [17:44] small business, like a hairdresser, for [17:46] example. I'm making it up. Or like a [17:48] massage therapist. So, the same concept [17:51] applies to everyone. And so, I'm on the [17:54] loyalty program for mine. And it says, [17:56] "This is this is interesting. Look, the [17:58] loyalty program is live. Here's what was [18:00] built. A pre-enrollment view pitch with [18:03] the three program benefits." M. [18:06] >> So, it's given me a value prop, a [18:08] preview stamp card showing seven out of [18:10] 10 and a simple sign up form, just a [18:13] name, no account needed. So, then [18:15] there's the dashboard post enrollment [18:17] live stamp card that animates as stamps [18:19] are added. So, you log in. So, this is [18:22] like you're logged into the system and [18:24] so you you understand and then you've [18:26] got four reward tiers. So, it's given me [18:28] regular, which is free coffee at 10, [18:30] sorry, category. That's 25 cups. You get [18:32] a free pastry and a coffee bonus pastry. [18:35] Oh, nice. [18:36] >> Then devoted a monthly free copy. So [18:39] that's I have to test that. That means [18:40] you're having more than a cup of coffee [18:42] a day. I don't know if that works. [18:45] >> But you know what I like about this is [18:46] that if you had the thought of starting [18:49] a loyalty program, it actually mapped [18:51] out a full program. [18:53] >> But it's an ideation process, is it? So [18:55] it actually gives you that ability. It's [18:56] not even it's not even the technical [18:58] barrier, but it's actually the smarts of [19:00] here's my idea. if I were to put this [19:04] into practice, how would I do it? So, it [19:07] gives you the the opportunity to [19:09] validate what you're thinking. And if [19:11] you don't like it, obviously you can [19:12] pressure test it and go back and forth [19:14] and get it to where you want it to be, [19:16] but it you're not alone. Do you know [19:19] what I mean? I think that's that's the [19:20] um the most beneficial part of it. So, [19:23] I'm going to go back and have a little [19:26] bit more of a chat around this. I'm [19:28] really enjoying this more than I thought [19:30] I would. This is kind of addictive. [19:33] >> It is [19:34] >> cuz I want to build more. I want to have [19:36] it functioning end to end. [19:38] >> But is it because you're empowered at [19:39] every step? So each step is like an [19:41] unlock. I mean, you know, when you when [19:42] everyone's like, "Oh, I'm vibe coding. [19:44] This is vibe coding." Like guy, you can, [19:46] you know, I I feel like going to [19:48] everyone doing this and this is vibe [19:49] coding. You don't have to you don't have [19:52] to be super technical to do it. That's a [19:55] realization. I think it's addictive [19:57] because I'm building something that in a [19:59] normal worldly life scenario I would not [20:02] be able to build this. [20:03] >> Can we I'm just going to pause for a [20:05] second. [20:06] >> So I asked it to do a confirmation email [20:09] on loyalty sign up and then it says [20:11] before I write the code what email [20:14] address should the confirmation go to? [20:16] The email will be sent to the person who [20:17] signs up their own address. When you [20:19] think about brand, [20:22] what I see is a lot of people, they come [20:24] in with a with a Gmail address. [20:26] >> Yeah. [20:26] >> And that's not a super convincing way to [20:31] >> talk with our customers. So there you'd [20:33] really need to have your um domain name [20:36] registered. And then you'd need to have [20:38] your info at or hello at Soho Brew. Is [20:43] it Soho Brew? Yes, Brew. um trademarking [20:46] my name [20:48] >> without having checked whether [20:49] >> this is it. [20:49] >> Um so that you do close the loop on the [20:52] trust uh the trustworthy relationship or [20:55] the you know building trust with the [20:56] with the audience. So they would be two [20:58] additional products that you'd pull into [21:00] this ecosystem that you have to make [21:03] sure that you're engaging with your [21:05] customers in the right way. Do that. I'm [21:06] going to go I haven't finished my [21:08] loyalty. I went off track a little bit. [21:09] So [21:10] >> do you want to share your screen and [21:11] show us your loyalty? [21:13] >> Take over. Oh, actually it's taken out a [21:16] stand workers. Oh, okay. So I But how do [21:20] I verify? So I've actually asked it a [21:22] qualifying question. Now what are you [21:24] asking it about? It says that the [21:27] customer So there's a virtual when you [21:28] log in. There's a virtual loyalty card, [21:32] a digital loyalty card, but I can just [21:34] add a stamp myself. I've asked it the [21:36] question on how does the customer verify [21:38] that they have purchased and received a [21:40] stamp. So right now the stamp is just [21:42] selfs served. the demo button. So, [21:44] option A, staff issued QR code, which is [21:47] what I had in mind, but I didn't have to [21:49] put put it in there because it's [21:50] defaulted to the QR code. So, customer [21:53] shows a loyalty page at the counter. [21:55] Staff scans a QR code and adds a stamp. [21:58] Simple, no customer account needed. [22:00] Purchase code after ordering a unique [22:02] code appears on the confirmation page. [22:04] Honest system and staff dashboard staff [22:06] add stamps manually from password. So, [22:09] this is where I'm getting into more of a [22:10] validation discussion around how I want [22:12] to build the program, which I think is [22:14] very interesting because yours went [22:16] directly to sign up, confirmation, and [22:20] contact, whereas mine is actually [22:22] focused more on here is the the [22:25] mechanism of recording the stance. So, [22:27] it's automatically understood what I'm [22:29] saying. Here's the full picture of what [22:31] you actually need. So, it's telling me [22:33] my two scenarios back. The most [22:35] practical setup is a staff facing [22:37] dashboard that handled both. Search by [22:39] name, redeem order code, password [22:42] protected. So this means a loyalty card [22:44] needs to live in a database rather than [22:45] just browser memory. Yes. [22:47] >> Okay. So now we're building the full [22:49] system. [22:50] >> Exactly. [22:50] >> So I've ended up where you are. [22:52] >> Yeah. [22:52] >> But it's educated me more in the [22:54] process. [22:55] >> I I was like, did you just build a front [22:57] end for the loyalty program or does it [22:59] really work after? [23:00] >> I was like, how can I make it work? [23:02] >> Yeah. So it said like you need this and [23:04] that and this to make it work. [23:05] >> But what I like here is you've gone with [23:07] a more technical specification in terms [23:09] of your question on how do I make it [23:11] work. I've taken the real life example [23:14] and said but I have more people coming [23:16] physically into my store. I'm not [23:18] entirely sure if this works. So I've [23:19] taken a very low tech approach. [23:22] >> Yeah. Yours I mean I wouldn't say yours [23:24] is super high-tech but you've taken more [23:26] of a okay does this actually work in [23:27] practice and it can. This is where [23:28] people actually need to go and like add [23:30] the prompts and ask it questions and get [23:32] it to work. [23:33] >> You have to test everything. I think [23:34] that's what is underestimated just [23:36] because you can see [23:39] it's a little bit what like when [23:40] somebody used to buy a domain name, [23:42] they'd sometimes I've seen examples like [23:44] this. They be like, "Oh, but I thought I [23:46] had a website with it." No, no, you have [23:47] to build the website. It's like you have [23:49] a mobile number, but you need to buy [23:51] data and a phone to make it work and [23:53] function. So, I think that that idea of [23:55] you have to go in and test it. you have [23:57] to build it beyond just the mockup, [23:59] beyond just the prototype is something [24:02] that many people may not understand. [24:04] Like you have to you have to do a [24:06] quality check of the entire flow to make [24:08] sure it does work. [24:09] >> Yeah. Um I want to show you one thing. [24:11] So right now I asked it to like send a [24:14] confirmation email on the loyalty signup [24:16] and then it says okay I now you need the [24:19] email and says like I have to have a [24:22] secret which is like in the API [24:24] integration. I'll need an SMTP setup to [24:27] send emails. So, let me request the [24:29] credentials. And that might seem like a [24:31] little bit scary to people, but you can [24:32] ask it what this means and so on. So, I [24:34] did [24:35] >> just pause and explain. I don't I just [24:38] jumped out of the I just like jumped out [24:40] the window cuz I don't understand what [24:41] you [24:42] >> So, basically what it says, what I'm [24:43] asking you to do is that every time [24:45] someone signs up to my loyalty program, [24:47] I want to send them an email just like [24:49] any other like website that you sign up [24:52] for. Uh, thank you so much for signing [24:53] up to your loyalty program. Some details [24:55] about the program, but you need to send [24:57] it from an email. And now what it's [24:59] >> as your home email basically. Exactly. [25:01] It's your business email. [25:02] >> Yeah. In this use case here, what it's [25:05] asking me to do is I need to integrate [25:08] or connect I'm going to use like an [25:09] easier word, connect my email to this [25:12] website to to the app that I'm building [25:15] cuz this is basically a web app to the [25:17] app that I'm building. So that every [25:19] time someone signs up for my loyalty [25:21] program, it triggers that email from my [25:23] email and then and then that person can [25:26] get back to my email and like write a [25:28] response if they want to get back to me. [25:29] So what it's asking me to do essentially [25:31] is getting a value which is the API key [25:34] value. And then if I go down here, I [25:36] said I'm unable to add it cuz I thought [25:38] all I needed to add was my email. But [25:41] now it's asking me, do you have an email [25:43] provider that you'd like to use? and [25:45] it's giving me like step by step do [25:46] this, do that, get the key and then fill [25:49] in the secret and just so I show [25:51] everyone what where the secret is or how [25:54] to go there. So basically those like [25:56] three dots when you click on them you go [25:59] to secrets and there you go. This is [26:01] like your secret management where you [26:03] can like add a new secret, give it a key [26:05] name and then you enter the secret [26:08] value. Um, and let me go to the database [26:11] really quickly cuz I did ask it to [26:12] create a data base for me. So, this is [26:15] it. So, it has a table here. Uh, I don't [26:17] have any data in my table yet. But once [26:19] it starts populating the data, I can add [26:22] rows, export it, and so on. So, again, [26:25] goes back to the idea of people just [26:27] want to add a prompt in and do it like [26:29] magically appears. But you have to do [26:31] some work and you have to learn those [26:33] things. It's not so hard. that actually [26:35] tells you exactly what to do. What are [26:37] the steps to do, which is it which is [26:39] what it's telling me here now. Uh I'm [26:41] just going to publish it and then I can [26:43] walk you through my website and I would [26:45] love to see your website. [26:47] >> But just seeing how indepth you can go [26:50] with your website [26:52] >> is in like for me this is something that [26:55] I would never be able to do honestly [26:57] without without this tool. [26:58] >> Yeah. No, same. And I if I look at it [27:00] over here, I suppose it's just it's the [27:02] same like both of both of us got um not [27:05] stuck. Both of us paused and spent the [27:08] most time on the loyalty program, which [27:10] shows that functionality is the one [27:12] that's the hardest to build because it's [27:13] dependent on integrations. And so here [27:16] for mine, I've gone down the route of [27:19] integrating the walk-in customers and [27:21] the online. So I've taken the offline [27:23] and the online together. And so it's [27:24] given me the loyalty to enroll for the [27:27] program. and then it's integrated it [27:30] with a staff pin. So for online orders [27:33] that goes in automatically. Um but the [27:36] staff they have to switch to the redeem [27:38] order code. So it's given me I'm not [27:39] going to go into detail. We can you know [27:41] when we run through it we can have a [27:42] look at it. But the staff basically have [27:45] their own dashboard where they access it [27:47] with a pin and then they're able to [27:49] verify the order [27:51] >> and then the customer gets the points. [27:53] So it's creating that loop to say what [27:55] is your objective? This is how we have [27:58] to go about designing the integration [28:02] and the the flow to be able to fulfill [28:04] that. And I don't think I'm I'm [28:06] completely done with this, but again, [28:07] for the sake of the demo, let's have a [28:09] look at the the final website. [28:10] >> Do you want to start? I want to see what [28:11] you built. And publishing is really [28:12] easy. [28:13] >> It's super easy. So, I would have I [28:16] would have edited the copy that the the [28:18] color scheme cuz I I think it fades [28:20] here. But it's very easy to do. So, I'm [28:22] not going to spend time. But here, I can [28:24] view the menu. So, it takes me down to [28:25] the menu. I've got the price list [28:27] already. It's taken local currency, [28:29] which I love. So, because I've said it's [28:31] based in the UAE and Dubai, it's already [28:33] taken it. This it's taken my the flat [28:36] white, the Australian classic. So, I [28:37] also because I spent time in my initial [28:41] prompt saying it's an Australian theme, [28:43] you know, Australian copy, we know our [28:45] copy really well. It's taken that and [28:46] integrated it into the text. [28:48] >> I love the text also that like the [28:50] explanation so like velvety micro form [28:53] over a double. [28:55] Yeah. And so copy copy is [28:57] >> copy is but if I wanted to have it less [29:00] descriptive then I could say make it [29:02] less like make it dry like make it more [29:04] dry cont uh content or copy. And so [29:07] here's my order. I did test this before. [29:10] I'm not sure if I was showing my screen. [29:12] So I'll just do it here quickly. Yeah, I [29:13] wanted a flat white. I add the flat [29:16] white. I say Selena [29:19] Bieber as soon as possible. I need that [29:22] pickme up. I want it with coconut milk [29:25] or I think in Australia they call it zy [29:27] milk. So I place my order and look here [29:30] actually after all those optimizations [29:32] loyalty stamp go show this code to staff [29:34] when you collect your order to earn a [29:36] stamp on your loyalty card. It says code [29:38] unavailable. So there's still a break in [29:41] my system that I need to you know work [29:43] with it a little bit more but it's [29:44] already integrating the updates that [29:46] I've done. And then if I go to the [29:47] loyalty system this is where I can join [29:50] it. But you can see it sets the rules [29:52] and this is just a demo. So then I would [29:54] say Selena Bieber and then I would say [29:57] join start earning stamps. This is where [30:00] there's an error because I haven't [30:01] completed the database integration. So [30:04] this is where your example will show us [30:06] how to do it. Um but also then it allows [30:08] me to find my card. So that you have a [30:11] login. So this is basically allowing the [30:13] customers it's it's a website plus a web [30:15] app in a way where you then have that [30:17] login system. a customer can go and [30:20] match their I mean on any e-commerce [30:22] website basically you can go in and see [30:24] what you've purchased your purchase [30:25] history but the fulfillment of the [30:27] payment is then in store right now but [30:29] then we can integrate online so there [30:31] are stages that I would go through [30:33] specifically for this site considering [30:35] we've been building for less than an [30:36] hour and you know we've been back and [30:37] forth on the features and the loyalty [30:39] program I think this is I'm happy with [30:41] where we're at um and have big dreams [30:44] for where I can clicked on publish and [30:47] this is how it looks like. I just want [30:49] to show one quick thing before I go to [30:51] this site that I published. So over here [30:53] it says, "Hey there, soob brws.com is [30:56] available." So I can actually trade [30:59] Mark. Who knows? [31:01] >> You've landed on a new business idea and [31:03] you go for it. Nutty dream [31:08] did not take off, but you definitely [31:09] can. [31:10] >> So this is another thing. You can buy [31:11] your domain and everything in one place [31:13] and just like take it online. Uh, so [31:16] this is my project. [31:17] >> It's a very sleek side. [31:19] >> Wait until you see my video. So [31:21] basically, I love the that I said that [31:24] my I want to have like a hippie and [31:26] stylish theme. So it's like what we pour [31:29] the good stuff. [31:30] >> Like the copy. I really like the good [31:32] stuff. I do. [31:33] >> I think the copy and then you see [31:35] everything is like animated and then the [31:38] reward system is here. The burgundy [31:40] color is here. [31:41] >> I love the burgundy color. I know you [31:43] didn't mean to change, but that I think [31:46] you ended up in the right at the right [31:48] destination. [31:49] >> I think so, too. I'm going to show you [31:51] really quickly. This is the menu page [31:52] that I built out. I would do like some [31:54] changes and have like a picture here, [31:57] but it still looks good. You get to see [31:59] all the options. As I said, you can [32:01] change the pictures. I don't like that [32:03] it shows the same picture, but this is a [32:06] very, very easy like replacement that [32:08] you can do yourself. Now, the loyalty [32:10] program. I'm so excited to see this and [32:13] see what it's giving you. [32:14] >> So, what I did not do is I did not set [32:16] up this I did not set up this properly [32:18] cuz I did not go through the API [32:20] integration end to end. But what what [32:22] would happen here if I did do the [32:24] integration is that when I enter my [32:26] name, my email address and join for [32:28] free, I would instantly get an email. [32:31] And now this is how it works. So, I [32:33] think my loyalty program is a little bit [32:35] different. You sign up for free and what [32:37] I would do is like in my shop I would [32:39] have a little QR code and then get the [32:42] baristas to say like scan the QR code. [32:45] You would go to this page where you [32:47] would sign up through this and then you [32:49] would be like signed up. You start [32:51] earning points. So for every point uh [32:53] get one point for every dollar and then [32:56] you get to redeem the rewards and then [32:58] like you level up. So these are the [32:59] tears and I want [33:00] >> Oh, I love that they've I love this is [33:04] amaz Look at this creativity. [33:06] >> It just It's so good. [33:07] >> Did you prompt this at all? No. [33:09] >> No, no, no. It did this all in it on its [33:11] own. Like [33:12] >> seedling bloom and rooted. This is [33:14] beautiful. I love it. [33:15] >> Yeah. And then like for 199 points up to [33:19] from 0 to 199 points, I get a birthday [33:22] treat. I get a member only newsletter [33:24] which I really like. And then 200 to [33:27] $4.99. Um, you get like everything in [33:30] seedling, double points on Tuesdays, [33:33] early access to seasonal ideas. It's [33:35] giving you marketing ideas within the [33:37] build. [33:38] >> I know. Free size upgrade once a month. [33:40] I love that. [33:42] >> I'm sold. I would come for Brew to have [33:44] coffee and then layering in marketing on [33:46] top. [33:46] >> So, it's like a business plan plus a [33:48] marketing plan plus a loyalty program [33:50] all in one. [33:52] >> And I did not even think about it. Now [33:54] that we're viewing it, we realize that [33:56] it's all of these included in basically. [33:58] You have to watch the signals, people. [34:00] You have to know what is coming together [34:02] because there is a lot that it's it's a [34:04] growth story. It's giving you a plan for [34:06] growing your business. I love it. [34:09] >> And then like you're supposed to be able [34:11] to watch your points grow with every [34:13] sip. So you would have like the balance, [34:15] how many points do you have, how many [34:17] points until the next tier, and like [34:19] what are the available rewards. So I I [34:22] would love to build that end to end. [34:24] >> So I think your site wins to be honest. [34:27] Congratulations. I think I won this [34:30] round. But I just love the power of [34:33] where one prompt. So we started with the [34:35] same prompt. We went in two very [34:36] different directions. One arguably took [34:38] us to a better result than the other. [34:40] But I think the learnings along the way [34:42] in terms of what it can do, what it [34:44] can't do, and asking us to question our [34:47] assumptions about what we initially [34:48] start. You and I had no idea about what [34:50] we wanted to build. [34:51] >> So already now we've indexed on the [34:54] discussion of how do you actually design [34:56] an effective loyalty program. I think [34:58] this is what I'm going to think about [34:59] for the rest of today because I still [35:02] have questions about how a couple coffee [35:04] shops here they do a loyalty program. [35:06] You add the card to your Apple wallet. [35:09] They scan the QR. So there's a linked [35:11] system but it gives you that that seed. [35:14] So you you're very you have something up [35:16] and running that you can continue to [35:18] customize and perfect. I love it. [35:20] >> Me too. I think for like my my departing [35:23] words is that I'm going to be thinking [35:25] about this. I'm going to go back and try [35:26] to build it like end to end cuz I want [35:28] to test it. I want to see if it works. [35:30] Yeah. [35:31] >> How it works. And like it's going to [35:33] take time, but I'm also not building [35:35] something simple. and building a full on [35:37] loyalty program that you would usually [35:39] pay hundreds or thousands of dollars for [35:41] sure to have it integrated into your [35:43] business and into your site. [35:45] >> And I think this is actually a good [35:46] topic for the next discussion or build [35:48] or podcast that we do in terms of what [35:52] would you have previously paid an agency [35:55] for and how much [35:56] >> versus what you can do with Aero AI [35:58] builder, how long and look at that [36:01] comparison. Yeah, love that.  ## 📺 Video ![](https://www.youtube.com/watch?v=wY2mXyuwu8I) --- # 📝 Full Notes (Deep-Read Analysis) # 🏷️ Title: Can AI Build an Entire Business? We Tried It Live. ## 🧰 Action Inbox - [H1] Test end-to-end loyalty program integration with database and email — 45:30 #business-automation - [H2] Compare AI builder costs vs traditional agency development — 52:15 #cost-analysis - [H3] Implement staff dashboard with QR code verification system — 32:45 #loyalty-programs ## 🧭 Table of Contents - [[#📝 Summary]] - [[#📖 Main Content]] - [[#🗒️ Notes]] - [[#📚 References]] - [[#🗣️ Paraphrase]] - [[#📰 Newsletter]] - [[#✅ Action Items]] - [[#🧪 Hypotheses]] - [[#📊 Methodology]] - [[#❓ Questions]] - [[#🔮 Future Research]] - [[#💡 Implications]] - [[#🧠 Learning Styles Assessment]] - [[#📊 Accessibility Assessment]] - [[#🌍 DEIB Assessment]] - [[#✅ Logical Validity]] - [[#🛡️ DARVO Analysis]] - [[#🎯 Conclusion]] ### 🔗 References - [Arrow AI Builder](https://arrow.ai) - [Getty Images Integration](URL) ## 📝 Summary - Two developers simultaneously built complete coffee shop websites using Arrow AI Builder in 15 minutes - Both started with identical prompts but created vastly different solutions with unique features - The AI automatically generated complex loyalty programs, ordering systems, and databases without technical expertise - Arrow AI introduced error handling that doesn't charge credits when the AI causes mistakes - The demo revealed how AI can create full business plans including marketing strategies and growth frameworks ## 📖 Main Content This live demonstration showcases Arrow AI Builder's capability to generate complete business applications from simple prompts. Two developers, starting with the identical brief of building a coffee shop website with online ordering and loyalty programs, ended up with completely different implementations - one in Dubai with an Australian theme, another in New York with a "hippie and stylish" aesthetic. The AI builder automatically generated sophisticated features including multi-tier loyalty programs with creative naming (Seedling, Bloom, Rooted), automated email confirmations, staff dashboards, and integrated databases. Beyond just building websites, the tool created comprehensive business strategies with marketing plans, customer retention frameworks, and operational workflows. A key product update was revealed during the build: Arrow AI now detects when errors are caused by the AI agent versus user input, automatically fixing agent-generated errors without charging credits to maintain fairness for customers. ## 🗒️ Notes #### ⏱️ [00:00](https://youtu.be/wY2mXyuwu8I?t=0) — [H1] Starting the simultaneous build challenge - Key idea: Both developers begin with identical coffee shop prompt but expect different outcomes - Evidence: "Same started, same starting point, but completely different results" #### ⏱️ [02:15](https://youtu.be/wY2mXyuwu8I?t=135) — [H2] AI automatically generates business plans and naming - Key idea: Arrow AI creates comprehensive project plans including branding decisions - Evidence: "It went ahead and did it by itself. So, it's like homepage, full menu page, online ordering integration, loyalty program page" #### ⏱️ [05:30](https://youtu.be/wY2mXyuwu8I?t=330) — [H3] Localization and currency adaptation - Key idea: AI automatically adapts to geographic context in prompts - Evidence: "It's taken local currency, which I love. So, because I've said it's based in the UAE and Dubai" #### ⏱️ [08:45](https://youtu.be/wY2mXyuwu8I?t=525) — [H1] Error handling and credit fairness system - Key idea: New feature prevents charging customers for AI-generated errors - Evidence: "We will assume the cost of the credits there versus charging the customer for the credits" #### ⏱️ [12:30](https://youtu.be/wY2mXyuwu8I?t=750) — [H2] Direct editing capabilities preserve credits - Key idea: Manual text editing bypasses AI processing to save computational costs - Evidence: "You can go in and directly edit the text... you don't consume credits which is great" #### ⏱️ [18:20](https://youtu.be/wY2mXyuwu8I?t=1100) — [H1] Complex loyalty program generation - Key idea: AI creates sophisticated multi-tier loyalty systems with creative branding - Evidence: "Seedling, bloom and rooted... It's giving you marketing ideas within the build" #### ⏱️ [25:45](https://youtu.be/wY2mXyuwu8I?t=1545) — [H2] Database integration and technical architecture - Key idea: AI explains backend requirements and guides technical implementation - Evidence: "To make it truly functional, you need a backend API route to receive and store signups" #### ⏱️ [32:15](https://youtu.be/wY2mXyuwu8I?t=1935) — [H3] Staff dashboard and verification systems - Key idea: AI designs practical in-store verification workflows - Evidence: "Staff scans a QR code and adds a stamp. Simple, no customer account needed" #### ⏱️ [38:50](https://youtu.be/wY2mXyuwu8I?t=2330) — [H2] Email integration and business identity requirements - Key idea: Professional email setup becomes crucial for customer trust - Evidence: "You'd really need to have your info at or hello at Soho Brew" #### ⏱️ [45:30](https://youtu.be/wY2mXyuwu8I?t=2730) — [H1] End-to-end functionality testing reveals gaps - Key idea: AI-built interfaces require thorough testing to ensure full functionality - Evidence: "You have to go in and test it. you have to build it beyond just the mockup" #### ⏱️ [48:15](https://youtu.be/wY2mXyuwu8I?t=2895) — [H3] Domain registration and business launch capabilities - Key idea: Platform integrates domain purchasing for complete business setup - Evidence: "Hey there, soob brws.com is available... You can buy your domain and everything in one place" #### ⏱️ [52:15](https://youtu.be/wY2mXyuwu8I?t=3135) — [H1] Cost comparison with traditional development - Key idea: AI builder delivers agency-level results at fraction of traditional cost - Evidence: "Building a full on loyalty program that you would usually pay hundreds or thousands of dollars" ## 📚 References - [Arrow AI Builder](https://arrow.ai) - [Getty Images Integration](https://getty.com) - [SMTP Email Setup Documentation](URL) ## 🗣️ Paraphrase - AI builders can generate complete business applications including sophisticated loyalty programs and databases from simple prompts in minutes rather than weeks - The same starting prompt can produce vastly different solutions, demonstrating AI's creative interpretation capabilities and the importance of iterative refinement - Professional business deployment requires additional considerations like domain registration, email setup, and thorough functionality testing beyond the initial AI build ## 📰 Newsletter **AI Builders Are Revolutionizing Small Business Development** This week, we witnessed something remarkable: two developers built complete coffee shop businesses with ordering systems and loyalty programs in just 15 minutes using AI. What started as identical prompts became unique solutions - one Dubai-based with Australian flair, another New York-focused with creative branding. The standout feature? Arrow AI's new error handling system that detects when mistakes are the AI's fault, not yours, and fixes them without charging credits. Plus, the integrated database creation, email systems, and even domain purchasing mean you can go from idea to live business faster than ever. The practical takeaway: AI isn't just building websites anymore - it's creating comprehensive business strategies, marketing plans, and operational workflows that would typically cost thousands from agencies. Resources: [Arrow AI Builder](https://arrow.ai), [No-Code Development Guide](https://learn.wordpress.org/), [Small Business Automation](https://developer.wordpress.org/) ## ✅ Action Items #### DevRel - Developer advocacy: Create tutorials showing AI builder integration with existing WordPress workflows - Education: Develop workshop curricula for no-code business development using AI builders - Developer experience: Document API integration patterns for loyalty systems and databases - Community management: Facilitate discussions on AI development ethics and credit fairness #### OSPO - Policy & Governance: Establish guidelines for AI-generated code licensing and attribution - Software Inventory: Track AI builder dependencies and generated code provenance - Developer Support: Provide training on AI-assisted development best practices #### Community - Maintainers: Support integration between AI builders and WordPress core features - Contributors: Create plugins that bridge AI-generated sites with WordPress ecosystem - Ecosystem: Coordinate standards for AI-generated business applications #### WordPress - Core: Evaluate AI builder integration opportunities for WordPress.com and .org - Marketing: Develop narratives around WordPress as platform for AI-enhanced businesses - Documentation: Create guides for integrating AI-built applications with WordPress backends - Hosting: Establish performance guidelines for AI-generated high-traffic applications #### The WPCC - Programs: Fund development of open-source alternatives to proprietary AI builders - Projects: Support documentation of AI development best practices for small businesses ## 🧪 Hypotheses - AI builders can reduce small business website development costs by 80-90% compared to traditional agencies - Identical AI prompts will produce functionally similar but aesthetically diverse results across multiple runs - Non-technical users can successfully deploy complex business applications using conversational AI interfaces - AI-generated business applications require human oversight for security and regulatory compliance ### 📊 Methodology - Comparative cost analysis between AI builder outputs and equivalent agency quotes - A/B testing of identical prompts across multiple AI platforms and sessions - User experience studies with non-technical small business owners using AI builders - Security audit framework for AI-generated business applications ### ❓ Questions - How do AI builders handle industry-specific regulations like PCI compliance for payment processing? - What happens to AI-generated businesses when the underlying platform changes pricing or features? - Can AI builders scale to enterprise-level complexity while maintaining the simplicity demonstrated? ### 🔮 Future Research - Integration patterns between AI builders and established CMS platforms like WordPress - Governance frameworks for AI-generated business applications in regulated industries - Long-term maintenance and evolution strategies for AI-built applications ### 💡 Implications - Practice: Small businesses can now deploy sophisticated digital infrastructure without technical teams - Policy: Regulatory frameworks need updating for AI-generated business applications and data handling - Security: AI-built applications require new audit approaches for generated code and database schemas - Education: Business education must include AI-assisted development as core entrepreneurship skill - Funding: Venture capital and small business lending models may shift toward AI-enabled rapid prototyping ### 🧠 Learning Styles Assessment Reading level estimate: 8th Grade (Flesch-Kincaid) #### 👀 Visual learners - Present: Screen recordings of live building process, UI mockups, color scheme demonstrations - Improve: Architecture diagrams showing database relationships, workflow visualizations #### 🔊 Aural learners - Present: Live commentary and discussion during building process - Improve: Podcast-style summary explaining technical concepts and business implications #### 🗣️ Verbal learners - Present: Conversational AI interaction examples, terminology explanations - Improve: Glossary of technical terms, business concept analogies #### 🧩 Kinesthetic learners - Present: Step-by-step building demonstration, hands-on prompt crafting - Improve: Interactive tutorial checklist, sandbox environment for testing #### 🧠 Logical learners - Present: Technical architecture explanations, cost-benefit analysis - Improve: Decision tree for choosing AI builder features, performance metrics dashboard #### 🧑‍🤝‍🧑 Social learners - Present: Collaborative building session, peer comparison of results - Improve: Community showcase of AI-built businesses, mentorship matching #### 🧘‍♀️ Solitary learners - Present: Individual building challenges, self-directed exploration - Improve: Personal project reflection prompts, progress tracking worksheets #### 🌿 Naturalistic learners - Present: Business ecosystem thinking, organic growth strategies - Improve: Sustainability impact analysis, local business integration patterns ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Clear heading structure in AI builder interface, descriptive button labels - Improve: Alt text for generated images, ARIA labels for dynamic content #### Keyboard navigation - Present: Tab navigation through builder interface - Improve: Skip links for complex generated forms, focus management in dynamic content #### Contrast and readability - Present: High contrast in builder interface - Improve: Ensure AI-generated sites meet WCAG contrast requirements by default #### Text resizing and reflow - Present: Responsive layouts in generated sites - Improve: Test generated content at 200% zoom, prevent horizontal scrolling #### Media and captions - Present: Video demonstration with spoken audio - Improve: Auto-generated captions for technical explanations, audio descriptions #### Forms and interactions - Present: Clear form labels in loyalty program signup - Improve: Error handling accessibility in AI-generated forms, validation announcements ### 🌍 DEIB Assessment #### Cultural breadth - Present: Multiple geographic contexts (Dubai, New York), cultural business themes (Australian coffee culture) - Improve: Examples from diverse global business contexts, multi-language AI builder support #### Inclusive language - Present: Gender-neutral business terminology, accessible technical explanations - Improve: Avoid assumptions about technical background, provide multiple complexity levels #### Neurodiversity support - Present: Clear step-by-step process, visual demonstrations - Improve: Adjustable pacing controls, alternative input methods for AI prompts #### Representation - Present: Diverse geographic and cultural business examples - Improve: Include accessibility-focused business examples, underrepresented entrepreneur showcases #### Psychological safety - Present: Supportive tone around technical challenges, error normalization - Improve: Explicit encouragement for non-technical users, failure-positive learning framing ### ✅ Logical Validity #### Strong reasoning - Deductive strength: Clear cause-effect relationships between AI prompts and outputs demonstrated throughout - Inductive strength: Solid evidence from live demonstration supports broader claims about AI capabilities - Causal reasoning: Well-established connections between feature complexity and development time/cost - Analogical strength: Effective comparisons between AI development and traditional agency work #### Fallacies - Appeal to novelty: Potential overemphasis on AI being "new" rather than focusing on practical utility - Hasty generalization: Limited sample size (two builds) used to make broad claims about AI capabilities #### Tone of voice - Overall tone: Enthusiastic and exploratory, with genuine surprise at capabilities and honest assessment of limitations ### 🛡️ DARVO Analysis - Deny: No denial patterns present - Attack: No attacking behavior observed - Reverse victim/offender: No victim/offender dynamics - Emotional cues and rhetoric: Positive excitement about technology capabilities, honest acknowledgment of complexity ## 🎯 Conclusion AI builders like Arrow AI are democratizing business application development by enabling non-technical users to create sophisticated systems in minutes rather than months. The live demonstration proved that identical starting prompts can produce diverse, functional solutions including complex loyalty programs, databases, and payment systems. However, success requires understanding that AI generates starting points that need testing, integration, and business-specific customization. The most significant barrier to small business digitization may no longer be technical complexity or cost, but rather knowing what questions to ask and how to iterate effectively with AI tools. ## 🧱 Refactor Appendix ### 📚 Literature Split Pad Purpose: Stage CEE units that become Literature notes. Split level: H3 ### [H1] AI builders reduce technical barriers for small business development > [!quote] [Can AI Build an Entire Business? We Tried It Live.](https://www.youtube.com/watch?v=wY2mXyuwu8I) by ArrowHQ via YouTube. Published 2024. Accessed 2024-12-28. > "People who wouldn't I mean I wouldn't have been able to do this... it would have taken time and reading and understanding." Demonstrates how AI eliminates traditional technical prerequisites for business application development. [view highlight](https://youtu.be/wY2mXyuwu --- # 📄 Full Transcript *Auto-generated from YouTube captions · timestamps roughly every 30 seconds · [watch on YouTube](https://www.youtube.com/watch?v=wY2mXyuwu8I)* **[00:00]** I read one hook, you read another, and it's okay. >> It's not the same thing though. So, you want you want to do it like you want to read the book first. >> Yeah. We picked one business idea. We're both building it live at the same time, talking through every prompt as we go. Same starting point, but completely different results. Let's do it. >> Okay. Let's go. Let's see where we end up. >> My suggestion is we build a neighborhood coffee shop has an online ordering with online ordering and a loyalty program. Let's **[00:31]** start. I have Arrow Builder open, so I'm going to start by sharing my screen. Yours is in Arabic? >> Yeah. >> Yeah, because I'm going to be testing Arabic today. It's going live. >> in Arabic mine can Is this the You just need to click on it in the navigation. All right, we're back. So, we said So, we said a website for a local coffee shop. >> Yeah. >> With a online ordering system and a loyalty program. Are we not Are we like going rogue on where it's where it's >> Yes. >> Yeah. >> Okay. I have a local coffee shop in the **[01:04]** Lakes, Dubai. Is there a We know our coffee. >> I'm building a local coffee shop in New York. >> Wow. That's my That's like my my dream location for a coffee shop. Who knows? I think it's going to come up with like a very local authentic I mean, we need one here. Maybe it'll take me to a new adventure. >> Exactly. >> It's a local coffee Start of a business. >> Yeah, it could be. We're an Australian that would He got >> loyalty program. There's a loyalty program. >> I know I'm going to make a dessert. She's going straight to desserts. >> Of course. You've gone past the coffee. >> The one thing that I know how to do the **[01:35]** best. Okay. I think I'm ready. Let's I'm going to give us like 15 minutes to build and then >> I'm done yet. I've like literally gone into the detail of the coffee. No, you start. Okay. I'm taking time with my prompt. >> It's 1:33 now. >> Okay. >> What's the coffee shop called? >> So, Lena is still writing your prompt. >> Yeah, I am. I'm not finished yet. >> I'm so proud of you. So, mine asked me a few questions and I created like a plan for me. >> Okay. Oh, you did you ask it to create a plan? >> No, no, no. It It went ahead and did it by itself. So it's like homepage, full **[02:06]** menu page, online ordering integration, loyalty program page, about and a story page. >> Mine's still thinking. It's named at the Lakes Coffee. >> The Lakes Coffee? >> Yeah. >> So it came up with the name? >> It did, which I would might change later, but that's a small detail cuz I haven't committed to a name yet. >> Okay. >> We can chat the I I came >> I came up with the name Soho Brew. >> Okay. I wonder if this exists elsewhere. These are the things where >> I might run into a a real Soho Brew **[02:36]** coming after me. >> You may. All right, you realize when you try to register your domain name. Is it taken or isn't it taken? That is the question. >> You know, it's going to tell me right now. >> Mhm. >> And like area I built already has this like in the domain URL, it recommends a domain name that is available for me to buy. >> Oh, I like that. Let's try. >> So even when I ran into my error, what I said is like try again. >> Yeah. >> And it's just like build and rethink. >> Yeah. So I it it gave me a little plan and now it's serving me with coffee with images actually for the homepage build. **[03:08]** It's given me a color scheme and font, but still think I don't like the colors that it's given me that much. So I like them with pastel. >> Which colors did it give you? >> Like a brown color palette. Um so >> Mine is like brown with greens. >> I don't have it It's done brown, black, and like a rusty type color which I'm not >> Which font did it give you? >> Uh Cormorant Garamond plus Lato. >> Me, too. >> And I think this is where I find you probably have to add the specifics in the design **[03:39]** >> Yeah. >> to get what you want. >> If you like a certain font. >> Yeah, cuz like I know what I don't want as well. So like this color scheme, I said make it like a modern and this hygge feel, you know, that Swedish comfort, but I think it's gone to um it's over-corrected to index on the copy colors versus something more pastel. >> I asked for like a hippie and stylish. >> Oh, I'm curious to see what that comes up with. >> Yeah, me too. >> And you know what I realized, you know when you and I both did the International Women's Day, I'm going off **[04:09]** topic here, but we did the International Women's Day, it gave us both a black and pink. >> Yes. No, mine was green and I asked it to change it to pink. >> I think I did the same. Isn't it weird that we defaulted it? Like I think it gave me the option of green and pink and I wanted pink. >> Okay. Mine actually built something for me and I really, really like it. >> Really? >> Yeah. >> Oh, good for you. I'm happy. Oh, wow. It's just up. So it gave me a little preview that I didn't like that much and **[04:39]** then it went back and thought about it and I'm not sure if I hate the color scheme as much as I had originally thought I did. >> I'm not as passionate about the font. >> I've got chat with Arrow. It's saying, "Tell Arrow what you want to build or modify in your app. The AI will understand your request and make changes." Okay, thank you. >> is saying something else. >> it saying? >> What do you think of the design? >> It's more confident. Look at you. Yeah, do you think it's the the prompt that we entered? >> I don't know. Something that was recently launched and now I'm seeing it and I'm so happy about it was that there **[05:09]** was a runtime error and then the chat prompted itself to fix and analyze the runtime error and then it says completed for free. >> Oh, that's amazing. >> It corrected the error and I was not credits were not charged. >> And this is I think it's important to pause on this one where, you know, technology will have errors. It's like when you have a bug on your phone and you have to restart or you have to restart your laptop. What we're really trying to do is recognize when Arrow AI builder has an error in credit or what is the cause of the error. So it's not a **[05:41]** user-generated error, it's a tool, it's an agent-generated error and so we will assume the cost of the credits there versus charging the customer for the credits. And so, that's now implemented and live. So, it's just a check mechanism to ensure that we continue to allow like fairness and build for the customer. So, we'll own the responsibility for uh for a generated by the agent and let you to continue to build as desired with the apps. But, it won't it won't fix your, you know, quality problems. You **[06:11]** still have to go and fix that yourself. >> Yeah. And like kudos to the product manager >> For sure. >> came up with this idea cuz we we have seen a lot of feedback from consumption of credits and it was a really thought-through process of like how can we fix this? >> I mean, it's it's a true example of putting the customer. I think it's two things. It's one going deeply into the product and understanding where the product is not meeting the bar in terms of there are certain tasks that need to be iterated. And again, like technology develops on a constant basis, you know, **[06:42]** especially in this agentic age. But, at the same time to the product manager who said, "Hey, this is an opportunity for us to do what's right for the customer." It was a very It was a very conscious process to say, "Okay, let's own accountability for what our tool does, what our agent does." And um and make sure that the customer isn't penalized for it. So, it's a proud moment. Like I feel I feel very proud to be on a >> Thank you, Josh. >> Yeah, thanks, Josh. Shout out. >> I will make some changes to my website now. >> asked me So, it's actually for me it's gone and said, "What do you think of the **[07:13]** design?" And rather than going and saying I don't like it, I'm first going to play with the color scheme and then I'm going to adjust the >> Oh, I'm going to do the same thing. I'm going to change my browns to burgundy. >> Did you get Maybe you would have liked my design better. >> Does yours have like burgundy in it? >> this like rusty orange. Do you think it's it's triggered by the copy? >> I think it's triggered by the vertical that we're building for. >> And probably it goes and finds like best colors that fit this vertical. >> So, I've I've gone in and I've said, "I don't like the colors. I want turquoise and greens across the page." And that's **[07:44]** what you did. >> Mine is like I said change the brown to burgundy. But what I meant to say was change the other burgundy. >> Can you You have to undo it now. >> I can undo it. I can actually go back to >> the previous >> to the backup and change it, but I like what I came up with, so I'm not hating it. >> Sometimes the right road is the wrong road. >> Yeah. Well, I love it. Now it's like Mine is like burgundy, green, and mint. >> Interesting. >> Yeah. >> We're still I'll rewrite the page with the new turquoise and cream palette **[08:15]** brown. Update the header. >> Please do. >> I talk a lot to the agent when I build. >> Yeah. >> Yeah, I feel like it needs to be more personable. >> I usually say like do this, do that. Sometimes when it's like not doing what I want, I say please. >> I type in please. I know it's a um it's a no-no because it uses credits and compute, and so in the grand scheme of environment, it's not a great thing. But I still feel like I don't want it to be death to politeness of humankind, so I feel like I still need to you know **[08:47]** invest in the etiquette that we've been taught. But don't do what I do because you will eat your credits. >> I just changed my font. So I I asked it asked it to change the font and give me some options. I did change it. Let me look at the new one. >> I think so I updated I don't think it's fully done. I think this color scheme does not work, but I'm not going to over invest time in this because we're already 10 minutes into the build and we said 15 minutes and I haven't I don't I don't have the online. I think I'm going to commit to writing I'm going to skip the font. I'm going to skip the **[09:18]** color scheme. I'm going to go straight to the build of the other pages so we actually see what it does. >> Okay. I'm I'm also building the other pages, but I'm I'm quite impressed by the speed. It's actually faster than what I expected. >> But I think it's important to point out here the speed when you look at the speed of different builders, I think where there's a quite a large difference is the detail at which it's trying to make the change. So, mine's just reloading now and it's actually redone the whole page in the new color palette. **[09:48]** >> Mhm. >> Um so, it took a little bit of time, but the output I'm just going to accept it. I don't think I made the right color choice, but I love this section here. It's super clean. Change the font cuz I just don't like this font. I wouldn't choose this font and not because I don't think it looks nice here, but I always go to physical print. So, let's say this is a real physical shop. Mhm. This kind of font I'm always a little bit iffy on if I have to print anything. It may not show up. I like more of the the clean and modern fonts just because, you know, whether you team it up with a logo or text, they're generally a little bit **[10:19]** more flexible in in output. Don't like these pictures. So, I would go back and change these into much more gourmet looking fresh veggies. >> Mhm. >> But I like this little picture down here. Every cup counts. Oh, this is so so even on the home page can do I'm curious to see what your loyalty looks like, but it's got a little loyalty program. Every cup counts. So, earn a stamp with every coffee, hit 10 and get your next one. No app required, just your a good habit. I'd asked it to do a point system, but **[10:49]** it's giving me a stamp system. So, it's default. So, there's a little bit of a disconnect, but it's taken my prompt to the 25, 50, and 100 cup cups cuz I'd said you hit like you get rewards at specific milestones. So, then it said you get bonus on a on top of the additional cup and free coffee, you get more on top of it. So, it's already giving me new ideas in terms of what I could do, but I'm going to Oh, but it must have tweaked the shade of the turquoise. You heard me, didn't you? You **[11:20]** heard me. Maybe I'll I'll have to call it something. >> Okay. So, right now I did build >> Let me have a look at yours. >> I did build the menu page. So, let me show you. This is what my This is what I the >> I love the colors. >> So beautiful. >> It's so nice. I'm slightly jealous. I feel like because you put matcha in the mix, it took the But, we've got the same picture though here. >> The the this one? >> yeah. >> The loyalty program, I gave it no. I love the burgundy. >> It looks lovely. **[11:51]** >> And then the menu is actually really clean. And when I when we were talking about changing the text without um without it without using credits, basically this is it. When you see this like little star over here, you click on it. >> And then you can go and edit it directly. >> I can remove this endlessly. >> And so, this is where that's actually really cool. So, people who are used to building in a template format, they can go in and directly edit the text. Yeah. So, if you have, you know, an existing website or anything that you have prepared, going back to the discussion **[12:23]** previously that we had around planning the content, you can actually take it and copy and paste it here. Cuz in this scenario, I would assume that you know what's on your menu. So, you can go in and just directly explain what that drink or what that beverage is to make sure that it looks truly, you know, the coffee type, the beans, etc. And you don't consume credits, which is great. So, you can be very smart about how you optimize your usage, but also you can make the builder as simple or as complex as you want to based on what you're **[12:53]** comfortable with. >> Yeah, and I think what the other thing is that like changing images can be so tricky with the agent cuz I tried it. Sometimes like drives me crazy. When you hover over the picture, click on replace, you can either replace it with one of the Getty images that we have, or you upload an image that and change it with it. >> easy. It's really simple. So, you don't have to And that saves you from having to write in a a really exact prompt to say this is exactly where I want you to take this **[13:24]** picture and put it in this placement on the page. You can you can get to your target position and your target image and put it in there. Yeah, I think that's great. And it's it's also good you don't actually have to have the images. You can go in and use Getty, which makes it You don't need to have anything prepared. You just need to know how to utilize it. So, it's done order ahead pick up ready. So, it's doing the ordering. It's quite detailed. I think we Even though it seems very simple, it's actually quite a detailed **[13:54]** task for the um for the agent to go out. So, you can see we've gone to On this one, we're building out the um oops, I've got an error now, so it's fixing it. Please analyze and fix this runtime error. >> I'm going to do like a big change now. I'm changing my header image to be a video. >> Wow. >> Cuz I enjoy No, I enjoy websites that have like >> Interrupt me. Yeah. You're such a next generation. I'm like, "Donna, don't overwhelm me." So, while you do that, I'm going to go **[14:25]** and check out my ordering system, which is built out. So, here, I actually love this. So, I have an ordering system. It's taken my queue on cortados and piccolos. I have everything except a double espresso. But, let's have a piccolo. I will go back >> We removed the matcha. >> the matcha. We don't serve matcha. And then it's given me a little option on the baked goods. But, this is I think again it goes back to the detail of when you have your menu, you can then input exactly what it that you offer. Bliss ball and seed bar. Okay, Ash is really **[14:56]** going for a bliss ball right now. So, here, piccolo and then I can say Bieber as soon as possible or I can say I would like to go for a coffee in 30 minutes. And then notes, I'm going to say coconut milk side of sugar. And then let's see what happens when I try place order. Thanks, Selena Bieber. The order is confirmed. And so, that is and then payment is at the counter. So, what we could do is we could go an extra mile and look at if I can do online payments so that the coffee shop doesn't prepare **[15:26]** the order and not get payment fulfillment. I think it's always safer to take the cash first, but that would be the next That would be another phase after I build the initial blueprint. But if I go back to home, then it takes me back to the home page. So, I'm going to look at the loyalty program now and see what it gives me. I'm going to let you screen share. Are you screen sharing now? >> I can start screen sharing. I'm so unhappy because my video is not working. Okay, so what I'm doing now is checking my loyalty program program to see if it **[15:57]** actually works. And I said, "Can you check if the loyalty program actually works?" And then it says, "To make it truly functional, you need a back-end API route to receive and store sign-ups. You need a database and a confirmation email sent on sign-ups." So, I'm like >> That's a lot. >> Yeah. >> He doesn't know how to do it. >> But it actually asks you how do you want to proceed and it has the has it planned out for you. So, you can start I I want to start by creating a database. And what is really really cool is that with the database, it's over here. Like it's **[16:29]** all inside the builder. You don't have to know how to create a database or you don't have to have it externally. I can show you now while it's being built. >> So, what you're effectively saying is the technical barrier doesn't exist anymore. >> Yeah. >> So, people who wouldn't I mean, I wouldn't have been able to do this. Maybe I would have. Let me not underestimate myself, but it would have taken time and reading and understanding to be able to get the database, integrate it, you know, capture leads, and then integrate it into the site as an order of full end-to-end loyalty and ordering system. **[16:59]** And then you want to have the engagement. You want to be able to contact those customers, you know, share them share information and promotions with them. But you the agent the builder does it for you now. You just have to give it the right instructions. So, that's it. That's incredible. >> Yeah. And when you think about it, you think, "Oh, I want to start a loyalty program." But did you really think it through of what it takes to have a loyalty program, all the steps that you need to have in your website to be able to get the loyalty program to work? And I think I like that it's telling me to have a loyalty program that is actually working, you need to do this and that **[17:30]** and this. And let's do it together. Let's build it together. >> Well, it allows you to understand the steps or the processes required to make it happen. And this isn't just for a coffee shop. I mean, this could be a community. This could be a you know, small business, like a hairdresser, I'm making it up, or like a massage therapist. So, the same concept applies to everyone. And so, I'm on the loyalty program for mine, and it says, "This is This is interesting. Look, the loyalty program is live. Here's what was built, **[18:00]** a pre-enrollment view, pitch with the three program benefits." >> Mhm. >> So, it's given me a value prop, a previous stamp card showing seven out of 10, and a simple sign-up form. Just a name, no account needed. And then there's the dashboard, post-enrollment, live stamp card that animates as stamps are added. So, you log in. So, this is like you're logged into the system, and so you you understand. And then you've got four rewards here. So, it's given me regular, which is free coffee at 10. Sorry, category. That's 25 cups, you get a free **[18:33]** pastry and a coffee. Bonus pastry. Then devoted, a monthly free coffee. So, that's I have to test that. That means you're having more than a cup of coffee a day. I don't know if that works. >> But, you know what I like about this is that if you had the thought of starting a loyalty program, it actually mapped out a full program of >> an ideation process. It's So, it actually gives you that ability. It's not even It's not even the technical barrier, but it's actually the smarts of Here's my idea. If I were to put this **[19:04]** into practice, how would I do it? So, it gives you the the opportunity to validate what you're thinking. And if you don't like it, obviously you can pressure test it and then go back and forth and get it to where you want it to be, but it you're not alone. Do you know what I mean? I think that's that's the the most beneficial part of it. So, I'm going to go back and have a little little bit more of a chat around this. >> I'm really enjoying this more than I thought I would. This is a kind of addictive. >> It is. **[19:34]** >> Cuz I want to build more. I want to have it functioning end to end. >> But is it because you're empowered at every step? So, that's like an unlock. I mean, you know, when you when everyone's like, "Oh, I'm bad at coding." This is bad coding. Like, guy, you can, you know, I I feel like going to everyone doing this. Like, "This is bad coding. You don't have to You don't have to be super technical to do it." Yeah, definitely that's a realization. >> I think it's addictive because I'm building something that in a normal worldly life scenario, I would not be able to build this. >> I'm just going to pause for a second. **[20:06]** >> So, I asked it to do a confirmation email on loyalty sign up. And then it says, "Before I write the code, what email address should the confirmation go to?" The email will be sent to the person who signs up, their own address. >> When you think about brand, what I see is a lot of people they come in with a with a Gmail address. Yeah. And that's not a super convincing way to >> No. >> talk with our customers. So, there you'd really need to have your domain name registered and then you'd **[20:37]** need to have your info at or hello at Soho Brew. Is it Soho Brew? Yes, Soho Brew. >> I'm trademarking my name. >> Without having checked Well, this is it. So, that you do close the loop on the trust, the trustworthy relationship or the, you know, building trust with the with the audience. And so, they would be two additional products that you'd pull into this ecosystem that you have to make sure that you're engaging with your customers in the right way. Do that. I'm going to go I haven't finished my **[21:08]** loyalty. I went off track a little bit. So >> Okay. Do you want to share your screen and show us your loyalty? >> over. Oh, actually it's taken Adam's stand. It's working. Oh, okay. So I But how do I verify? So I've actually asked it a qualifying question now. >> What are you asking it about? >> It says that the customer So there's a virtual when you log in there's a virtual loyalty card, a digital loyalty card, but I can just add a stamp myself. So I asked it a question on how does the customer verify that **[21:39]** they have purchased and received a stamp? So right now the stamp is just self-serve, the demo button. So option A, staff issued QR code, which is what I had in mind, but I didn't have to pop pop put it in there because it's defaulted to the QR code. So customer shows their loyalty page at the counter, staff scans the QR code and adds a stamp. Simple, no customer account needed. Purchase code after ordering, a unique code appears on the confirmation page. Honest customer and staff dashboard. Staff adds stamps manually from password pro So here this is where **[22:09]** I'm getting into more of a validation discussion around how I want to build a program, which I think is very interesting cuz yours went directly to sign-up confirmation and contact, whereas mine is actually focused more on here is the the mechanism of recording the stamps. So it's automatically understood what I'm saying. Here's the full picture of what you actually need. So it's telling me my two scenarios back. Most practical setup is a staff-facing dashboard that handles both. Search by name, redeem order code, **[22:41]** password protected. So this means a loyalty card needs to live in a database rather than just browser memory. >> Mhm. >> Okay. So now we're building the full system. >> Exactly. >> So I've ended up where you are. >> Yeah. >> But it's educated me more in the process. >> I I was like, "Did you just build a front end for the loyalty program, or does it really work?" >> Is it PR after? >> I was like, "How can I make it work?" >> Yeah. >> So it said, like you need this and that and this to make it work. >> But what I like here is you've gone with a more technical specification in terms of your question on how do I make it **[23:11]** work? I've taken a real-life example and said, but I have more people coming physically into my store. I'm not entirely sure if this works. So I've taken a very low-tech approach. >> Yeah. >> Um you're I mean, I wouldn't say yours is super high-tech, but you've taken more of a okay, does this actually work in practice and it >> And this is where people actually need to go and like add the prompts in ask a questions and get it to work. >> to test everything. I think that's what is underestimated. Just because you can see it's a little bit what like when somebody used to buy a domain name, they **[23:42]** they'd sometimes I've seen examples like this. They'd be like, oh, but I thought I had a website with it. No, no, you have to build the website. It's It's like you have a mobile number, but you need to buy data and a phone to make it work and function. So I think that that idea of you have to go in and test it, you have to build it beyond just the mock-up, beyond just the prototype is something that many people may not understand. Like you have to you have to do a quality check of the entire flow to make sure it does work. >> Yeah. Um I want to show you one thing. So right now I asked the tool like send **[24:14]** a confirmation email on the loyalty sign-up. And then it says, okay, I now you need the email and says like I have to have a secret, which is like in the API integration. I'll need an SMTP setup to send emails. So let me request the credentials. And that might seem like a little bit scary to people, but you can ask it what this means and so on. So I did try >> I just pull that and explain? I don't I just jumped out of the I just like jumped out the window cuz I don't understand what you're >> So basically what it says, what I'm asking you to do is that every time **[24:45]** someone signs up to my loyalty program, I want to send them an email. Just like any other like website that you sign up for. Uh thank you so much for signing up to your loyalty program, some details about the program. But you need to send it from an email. And now, what it's asking you is your home email, basically. Exactly. It's your business email. In this use case here, what it's asking me to do is I need to integrate or connect I'm going to use like an easier word, connect my email to this website to to do the app that I'm building cuz **[25:15]** this is basically a web app to the app that I'm building so that every time someone signs up for my loyalty program, it triggers that email from my email. And then and then that person can get back to my email and like write a response if they want to get back to me. So, what it's asking me to do essentially is getting a value, which is the API key value. And then, if I go down here, I said I'm unable to add it cuz I thought all I needed to add was my email. But, it now it's asking me, "Do you have an email provider that you'd like to use?" And it's giving me like **[25:45]** step-by-step, do this, do that, get the key, and then fill in the secret. And just so I show everyone where the secret is or how to go there. So, basically those like three dots, when you click on them, you go to secrets, and there you go. This is like your secret management where you can like add a new secret, give it a key name, and then you enter the secret value. Um And let me go to the database really quickly cuz I did ask it to create a database for me. So, this is it. So, it has a table here. Uh I don't have any **[26:18]** data in my table yet, but once it starts populating the data, I can add rows, export it, and so on. So, again, goes back to the idea of people just want to add a prompt in and do it like magically appears, but you have to do some work, and you have to learn those things. It's not so hard. It actually tells you exactly what to do, what are the steps to do, which is which is what it's telling me here now. Uh I'm just going to publish it, and then I can walk you through my website, and I would love to see your website. But, just seeing how in-depth you can go **[26:50]** with your website. Is it like for me this is something that I would never be able to do honestly without without this tool. >> Yeah. No, I say me and I if I look at it over here I suppose it's this it's the same like both of both of us got not stuck. Both of us paused and spent the most time on the loyalty program which shows that functionality is the one that's the hardest to build because it's dependent on integrations. And so here for mine I've gone down the route of integrating the walk-in customers and **[27:21]** the online. So I've taken the offline and the online together and so it's given me the loyalty to enroll for the program and then it's integrated it with a staff pin. So for online orders that goes in automatically but the staff they have to switch to the redeem order code. So it's given me I'm not going to go into detail we can you know when we run through it we can have a look at it but the staff basically have their own dashboard where they access it with a pin and then they're able to verify the order and then the customer gets the **[27:52]** points. So it's creating that loop to say what is your objective this is how we have to go about designing the integration and the the flow to be able to fulfill that. And I don't think I'm I'm completely done with this but again for the sake of the demo let's have a look at the final website. >> to start? I want to see what you built. And publishing is really easy. >> It's super easy. So I would have I would have edited the copy but the the color scheme cuz I I think it fades here but it's really easy to do so I'm not going to spend time. But here I can view the **[28:24]** menu so it takes me down to the menu. I've got the price list already. It's taken local currency which I love. So because I've said it's based in the UAE in Dubai it's already taken it. It's it's taken my the flat white the Australian classic so I've also because I spent time in my initial prompt saying it's an Australian theme you know Australian coffee we know our coffee really well it's taken that and integrated it into the text. >> I love the text also that it like the explanation. So like velvety microfoam over a double. **[28:54]** >> Yeah, it looks good. Okay. And so I >> Copy copy is >> But if I wanted to have it less descriptive, then I could say make it less make it dry. Like make it more dry font content or copy. And so here's my order. I did test this before. I'm not sure if I was showing my screen. So I'll just do it here quickly. Yeah, I wanted a flat white. I add the flat white. I say Selena Bieber as soon as possible. I need that pick-me-up. I want it with coconut milk. **[29:25]** Or I think in Australia they call it soy milk. So I place my order. And look here actually after all those optimizations, loyalty stamp it show this code to staff when you collect your order to earn a stamp on your loyalty card. It says code unavailable. So there's still a break in my system that I need to you know work with it a little bit more, but it's already integrating the updates that I've done. And then if I go to the loyalty system, this is where I can join it. But you can see it sets the rules and this is just a demo. So then I would say Selena Bieber. And then I would say **[29:57]** join, start earning stamps. This is where there's an error because I haven't completed the database integration. So this is where your example will show us how to do it. But also then it allows me to find my card. So that you have a login. So this is basically allowing the customers it's it's a website plus the web app in a way where you then have that login system. A customer can go and match their I mean on any e-commerce website basically. You can go in and see what you purchase, your purchase history. But the fulfillment of the payment is then in **[30:28]** store right now, but then we can integrate online. So there are stages that I would go through specifically for this site. Considering we've been building for less than an hour and you know we've been back and forth on the features and the loyalty program. I I this is I'm happy with where we're at. Um and I have big dreams for where I can like make >> clicked on publish and this is how it looks like. I just want to show one quick thing before I go to this site that I published. So over here it says, "Hey there, so whole brews.com is available." So I can actually trademark. **[31:00]** Who knows? You landed on a new business idea. >> And you go for it. Not your dream but my my late dream did not take off but yours definitely can. >> So this is another thing. You can buy your domain and everything in one place and just like take it online. So this is my project. >> It's a very sleek site. >> Wait until you see my video. So basically I love the that I said that my web I want to have like a hippie and stylish theme. So it's like what we pour, the good stuff. **[31:30]** >> I like the copy. I really like the good stuff. I do. >> I think the copy and then you see everything is like animated and then the reward system is here. The burgundy color is here. >> I love the burgundy color. I know you didn't mean to change but that I think you ended up in the right in the right destination. >> I think so too. I'm going to show you really quickly this is the menu page that I built out. I would do like some changes and have like a picture here but it still looks good. You get to see all the options. As I said you can change **[32:02]** the pictures. I don't like that it shows the same picture but this is a very very easy like replacement that you can do yourself. Now the loyalty program. So what I did not do is I did not set up this I did not set up this properly cuz I did not go through the API integration end to end but what what would happen here if I did do the integration is that when I enter my name, my email address and join for free I would instantly get an email. And now this is how it works. So **[32:33]** I think my loyalty program is a little a different. You sign up for free and what I would do is like in my shop, I would have a little QR code, and then get the baristas to say like scan the QR code, you would go to this page where you would sign up through this, and then you would be like signed up. You start earning points, so for every point uh get one point for every dollar, and then you get to redeem the rewards, and then you like you level up. So, these are the tiers, and I >> Oh, I love that they I loved This is **[33:04]** amazing. Look at this creativity. >> It just It's so good. >> Did you tell this at all? >> No. No, no, no. It did this all in it on its own. >> Like feeling gloomy and ruthless. This is visual. I love it. >> Yeah. And then like for 199 points up to from zero to 199 points, I get a birthday treat. I get a member-only newsletter, which I really like. And then 200 to 499, um you get like everything in Seedling, double points on Tuesdays, early access to seasonal **[33:34]** menus. >> So, ideas, it's giving you marketing ideas within the bill. >> I know. Free size upgrade once a month. I love that. >> I'm sold. I would come to Seedling to have coffee and then layering and marketing on top. >> So, it's like a business plan plus a marketing plan plus a loyalty program all in one. >> And I did not even think about it. >> Now that we're viewing it, we realize that it's all of these included in the basically. >> Yeah. >> You have to watch the signals, people. You have to know what is coming together because there is a lot that it's it's a growth story. **[34:05]** It's giving you a a plan for growing your business. I love it. >> And then like you're supposed to be able to watch your points grow with every sip, so you would have like the balance, how many points do you have, how many points until the next tier, and like what are the available rewards. So, I I would love to build that end to end. >> So, I think your site wins, to be honest. Uh congratulations. I think Otter won this round. But I just love the power of where one prompt. So we started with the same prompt. We went in **[34:36]** two very different directions. One arguably took us to a better result than the other, but I think the learnings along the way in terms of what it can do, what it can't do, and asking us to question our assumptions about what we initially start. You and I had no idea about what we wanted to build. So already now we've indexed on the discussion of how do you actually design an effective loyalty program? I think this is what I'm going to think about for the rest of today because I still have questions about how a couple coffee shops here, they do a loyalty program, **[35:06]** you add the card to your Apple wallet, they scan the QR. So there's a little system, but it gives you that that seed. So and you're you're very you have something up and running that you can continue to customize and perfect. I love it. >> Me too. I think like my my departing words is that I'm going to be thinking about this. I'm going to go back and try to build it like end-to-end cuz I want to test it. I want to see if it works. Yeah. How it works and like it's going to take time, but I'm also not building something simple. I'm building a full-on loyalty **[35:38]** program that you would usually pay hundreds or thousands of dollars to have it integrated into your business and into your site. >> this is actually a good topic for the next discussion or build or podcast that we do in terms of what would you previously paid an agency for and how much versus what you can do with Arrow AI builder and how long and look at that comparison. Yeah. Love that.