We Built an AI Team That Actually Feels Human

We Built an AI Team That Actually Feels Human 🔗 **Source:** [We Built an AI Team That Actually Feels Human](https://www.youtube.com/watch?v=C5DdWhw5LLM) > [!note]+ Document Notes > tags: [ai-design, character-design, product-development] > author: Vito Peleg > source: YouTube > reference: https://www.youtube.com/watch?v=C5DdWhw5LLM > date created: 2024-12-19 14:30 > > # 🏷️ Title: We Built an AI Team That Actually Feels Human >  ## 🧰 Action Inbox - [H1] Study creative audience expectations when designing AI interfaces — 02:15 #design - [H2] Research Picasso's cubism influence on character design — 08:45 #art-direction - [H3] Document character personality mapping to real team members — 06:30 #character-development ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) ### 🔗 References - [Original Video](https://www.youtube.com/watch?v=C5DdWhw5LLM) - [Picasso Cubism Reference](https://en.wikipedia.org/wiki/Cubism) ## 📝 Summary - A product team transformed generic AI robot characters into distinctive, personality-driven team members - Each AI agent represents a specific role (designer, SEO expert, project manager) with unique characteristics - The design process evolved from boring robots to Picasso-inspired cubist characters with backstories - Creative audiences require higher aesthetic standards than generic implementations - Real team member personalities served as inspiration for character development ## 📖 Main Content The video documents the evolution of AI character design from generic robots to distinctive team members with individual personalities. The creator emphasizes "digitizing human experiences" by presenting AI agents as actual team members rather than faceless bots. The design philosophy centers on serving creative audiences who expect higher aesthetic standards. Instead of typical robot icons, the team developed characters like Pixel (the frustrated designer), Index (the speedy SEO expert), and Clero (the professional project manager). The artistic approach draws heavily from Picasso's cubism, incorporating the concept of showing multiple perspectives simultaneously. Characters intentionally appear weathered and experienced, suggesting they've "been through the struggles" on behalf of users. ## 🗒️ Notes #### ⏱️ [00:00](https://youtu.be/C5DdWhw5LLM?t=0) — [H1] Initial robot designs were boring - Key idea: First versions were generic smiling robots in different colors - Evidence: "The first version were just like a boring robot... very much like the safe approach" #### ⏱️ [00:35](https://youtu.be/C5DdWhw5LLM?t=35) — [H2] Team member concept emergence - Key idea: AI agents should represent actual specialized team members - Evidence: "One is a designer, one is an SEO expert, one is the project manager" #### ⏱️ [01:15](https://youtu.be/C5DdWhw5LLM?t=75) — [H2] Digitizing human experiences philosophy - Key idea: Instead of replacing humans, AI should embody human team collaboration - Evidence: "If we are digitizing human experiences, who is actually sitting on the other side? It's an actual team member" #### ⏱️ [02:15](https://youtu.be/C5DdWhw5LLM?t=135) — [H1] Creative audience requirements - Key idea: Creative audiences demand higher aesthetic standards than generic segments - Evidence: "Our audience are creative people... the level of aesthetics that is required to serve this audience is different" #### ⏱️ [03:30](https://youtu.be/C5DdWhw5LLM?t=210) — [H3] Shape and form exploration - Key idea: Design process started with basic geometric shapes and evolved organically - Evidence: "Let's just try and start with the shape. Maybe the top bit here is the nose" #### ⏱️ [04:45](https://youtu.be/C5DdWhw5LLM?t=285) — [H3] Iterative design process documentation - Key idea: Design fidelity improved through rapid iteration and experimentation - Evidence: "You literally see the thinking process and how things shaped up" #### ⏱️ [05:30](https://youtu.be/C5DdWhw5LLM?t=330) — [H2] Character personality development - Key idea: Each character has distinct personality traits based on their role - Evidence: "I tried to give each one of them characteristics and personalities based on how they act and think" #### ⏱️ [06:30](https://youtu.be/C5DdWhw5LLM?t=390) — [H2] Real person inspiration - Key idea: Characters are caricatures of real people in similar roles - Evidence: "For most of them I have a person in mind... they're mocked after people that I know" #### ⏱️ [07:00](https://youtu.be/C5DdWhw5LLM?t=420) — [H3] Individual character traits - Key idea: Visual design elements reinforce each character's professional role - Evidence: "Lexi... words on her clothes... Index... indicate speed... Pixel is the frustrated designer" #### ⏱️ [08:00](https://youtu.be/C5DdWhw5LLM?t=480) — [H2] AI enhancement of 2D characters - Key idea: AI tools were used to evolve basic 2D designs into more sophisticated versions - Evidence: "I can just leverage AI to level them up... bringing those basic 2D characters into the game" #### ⏱️ [08:30](https://youtu.be/C5DdWhw5LLM?t=510) — [H3] Weathered experience concept - Key idea: Characters should appear experienced and battle-tested - Evidence: "These characters has seen things... that's why you have all this grime and dirt on the face" #### ⏱️ [08:45](https://youtu.be/C5DdWhw5LLM?t=525) — [H1] Picasso cubism influence - Key idea: Cubist art principles informed the character design approach - Evidence: "Picasso was a big influence... bringing the entire perspective into the front" #### ⏱️ [09:30](https://youtu.be/C5DdWhw5LLM?t=570) — [H2] Multiple perspective integration - Key idea: Characters show multiple angles simultaneously, like cubist paintings - Evidence: "You might see the ear that is on the back and you might see the eye that you can't see otherwise" ## 📚 References - [We Built an AI Team That Actually Feels Human](https://www.youtube.com/watch?v=C5DdWhw5LLM) - [Picasso and Cubism](https://www.pablopicasso.org/cubism.jsp) - [Character Design Principles](https://www.creativebloq.com/character-design) ## 🗣️ Paraphrase - AI interfaces don't have to be generic robots - they can embody human team member characteristics while maintaining their functional roles - Creative audiences expect higher design standards, requiring thoughtful aesthetic consideration rather than safe, expected solutions - Drawing inspiration from artistic movements like cubism can create unique visual approaches that stand out from typical AI implementations ## 📰 Newsletter **From Boring Bots to Picasso-Inspired AI Team Members** Ever wondered why most AI interfaces look like generic smiling robots? One design team decided to break the mold by creating AI agents that actually feel like human colleagues. Instead of faceless chatbots, they developed characters like Pixel (the perfectionist designer) and Index (the speed-focused SEO expert) - each with distinct personalities inspired by real team members. The secret? Treating AI as a way to "digitize human experiences" rather than replace them. **Key Takeaway**: Creative audiences expect better than cookie-cutter robot designs. By incorporating artistic influences (hello, Picasso!) and real personality traits, AI interfaces can become memorable and engaging. **Try This**: Next time you're designing an AI feature, ask "Who would actually be doing this job?" Then give that persona visual life. Resources: [Learn WordPress](https://learn.wordpress.org/), [Developer Docs](https://developer.wordpress.org/), [WP-CLI](https://wp-cli.org/) ## ✅ Action Items #### DevRel - Developer advocacy: Create character-driven documentation personas for different developer types - Education: Develop workshops on humanizing technical interfaces beyond generic bot designs - Developer experience: Research how personality-driven AI affects developer tool adoption - Developer marketing: Document case studies of successful character-based AI implementations - Community management: Survey creative communities about AI interface preferences #### Community - Contributors: Gather feedback on whether WordPress AI features should adopt character-driven approaches - Maintainers: Explore personality frameworks for different WordPress AI tools and assistants - Ecosystem: Study how character design affects user engagement across WordPress plugins #### WordPress - Design: Research character-based approaches for WordPress AI features and tools - Documentation: Consider persona-driven help documentation that matches user mental models - Community: Survey WordPress creative community about AI interface expectations and preferences ## 🧪 Hypotheses - Character-driven AI interfaces increase user engagement compared to generic robot designs - Creative audiences show higher satisfaction with aesthetically sophisticated AI implementations - Personality-based AI agents improve user trust and perceived competence - Cubist-inspired multi-perspective character design enhances memorability and brand recognition ### 📊 Methodology - A/B testing character-driven vs. generic AI interfaces measuring engagement time and task completion - User satisfaction surveys comparing aesthetic preferences across creative vs. non-creative user segments - Eye-tracking studies on character design elements to identify attention patterns and emotional responses ### ❓ Questions - How does character personality consistency affect long-term user relationships with AI tools? - What specific visual elements most effectively communicate AI agent competence and trustworthiness? - How can artistic influences be systematically applied to functional interface design without sacrificing usability? ### 🔮 Future Research - Cross-cultural studies on character design preferences in AI interfaces across global creative communities - Longitudinal studies tracking user attachment and engagement with personality-driven AI over extended usage periods - Investigation of artistic movement influences beyond cubism in contemporary AI interface design ### 💡 Implications - Practice: Design teams should invest in character development and artistic research for AI interfaces serving creative audiences - Policy: AI interface design guidelines should account for audience-specific aesthetic expectations rather than one-size-fits-all approaches - Education: Design curricula should incorporate art history and character development alongside technical AI interface principles - Funding: Budget allocation for AI projects should include character design research and artistic consultation ### 🧠 Learning Styles Assessment Reading level estimate: 8th Grade (Flesch-Kincaid) #### 👀 Visual learners - Present: Character design evolution sketches, before/after comparisons, artistic influence examples - Improve: Step-by-step design process diagrams, color theory applications, geometric shape breakdowns #### 🔊 Aural learners - Present: Video narration explaining design decisions and artistic influences - Improve: Audio commentary on character personality development, design critique discussions #### 🗣️ Verbal learners - Present: Clear character personality descriptions, artistic terminology definitions, design rationale explanations - Improve: Analogies between artistic movements and interface design, storytelling approaches to character development #### 🧩 Kinesthetic learners - Present: Iterative design process demonstration, hands-on character sketching examples - Improve: Interactive character creation exercises, physical art material experiments, design workshop templates #### 🧠 Logical learners - Present: Design decision frameworks, systematic character trait mapping, artistic principle applications - Improve: Decision trees for character design choices, measurable personality trait frameworks #### 🧑‍🤝‍🧑 Social learners - Present: Team collaboration insights, real person inspiration discussions - Improve: Character design critique sessions, collaborative personality development exercises #### 🧘‍♀️ Solitary learners - Present: Individual artistic exploration examples, personal design journey documentation - Improve: Self-reflection prompts on character preferences, independent art study guides #### 🌿 Naturalistic learners - Present: Organic design evolution process, natural character development progression - Improve: Biological analogies for character trait development, ecosystem thinking about AI team dynamics ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Character names and role descriptions clearly labeled - Improve: Detailed alt text for character visual elements, personality trait audio descriptions #### Keyboard navigation - Present: Video player controls accessible via keyboard - Improve: Character gallery navigation, design process step-through controls #### Contrast and readability - Present: High contrast character designs for visual clarity - Improve: Background color considerations for character visibility, text overlay contrast #### Text resizing and reflow - Present: Responsive character design examples - Improve: Scalable character elements, flexible layout accommodation #### Media and captions - Present: Video narration available - Improve: Design process step descriptions, artistic terminology captions #### Forms and interactions - Present: Basic video controls - Improve: Interactive character customization tools, design feedback forms ### 🌍 DEIB Assessment #### Cultural breadth - Present: References to French nationality in character design - Improve: Multi-cultural artistic influence examples, global character design patterns #### Inclusive language - Present: Gender pronouns used respectfully for characters - Improve: Non-binary character options, inclusive personality trait descriptions #### Neurodiversity support - Present: Visual and auditory learning approaches combined - Improve: Sensory-friendly character design options, attention-friendly presentation formats #### Representation - Present: Mix of character personalities and roles - Improve: Diverse character ethnicities, varied ability representations, inclusive team compositions #### Psychological safety - Present: Positive framing of creative process struggles - Improve: Validation of different artistic preferences, non-judgmental design critique frameworks ### ✅ Logical Validity #### Strong reasoning - Deductive strength: Creative audiences require higher aesthetics → Character design improves engagement (02:15) - Inductive strength: Multiple character iterations led to better results → Systematic iteration improves design (04:45) - Causal reasoning: Picasso influence → Multi-perspective character design → Enhanced memorability (08:45) - Analogical strength: Human team member qualities → AI character personalities → Improved user relationships (06:30) #### Fallacies - Appeal to authority: Picasso influence cited without demonstrating direct applicability to AI interface design - Hasty generalization: Creative audience preferences assumed universal without supporting data #### Tone of voice - Overall tone: Enthusiastic and creative, with personal artistic passion evident throughout. Conversational and accessible while maintaining professional design insights. ### 🛡️ DARVO Analysis - Deny: No denials or defensive positioning observed - Attack: No attacks on alternative approaches or critics present - Reverse victim/offender: No victim/offender dynamics identified - Emotional cues and rhetoric: Positive enthusiasm for creative problem-solving, collaborative tone throughout ## 🎯 Conclusion The evolution from generic AI robots to Picasso-inspired team member characters demonstrates that thoughtful design can transform user relationships with AI tools. By treating AI as a way to "digitize human experiences" rather than replace them, design teams can create more engaging and memorable interfaces. The key insight is understanding audience expectations - creative users demand higher aesthetic standards and benefit from personality-driven interactions. The systematic approach of mapping real team member traits to AI characters, combined with artistic influences like cubism, creates distinctive and functional design solutions. For practitioners, this suggests investing in character development research, artistic consultation, and iterative design processes when creating AI interfaces for creative audiences. The approach offers a template for moving beyond safe, expected solutions toward memorable and effective AI user experiences. ## 🧱 Refactor Appendix ### 📚 Literature Split Pad Purpose: Stage CEE units that become Literature notes. Split level: H3 ### [H1] Creative audiences require higher aesthetic AI standards > [!quote] We Built an AI Team That Actually Feels Human by Vito Peleg --- ## Full Transcript [00:00] the first version were just like a [00:01] boring robot. Okay, it's AI. It's going [00:04] to be a robot that is smiling. If we are [00:06] digitizing human experiences, who is [00:08] actually sitting on the other side? It's [00:09] an actual team member. One is a [00:11] designer, one is an SEO expert, one is [00:13] the project manager. Let's just try and [00:15] start with the shape. Maybe the top bit [00:17] here is the nose. Kind of threw in that [00:19] little antenna here that helps balance [00:22] the head. Index is in charge of SEO. I [00:24] wanted to indicate speed. Pixel is the [00:27] frustrated designer. Nothing is good. he [00:29] can see what we're doing wrong. Also, [00:31] thinking about this from an art [00:32] perspective, how can I make them be art [00:35] pieces in their own right, Picasso was a [00:37] big influence on how I was painting with [00:39] oils and things like this. [00:44] When we first started working on the AI [00:47] implementation, I started with that [00:48] mindset of just digitizing human [00:50] experiences, which has been the mantra [00:52] for quite a few years in terms of how [00:54] we're trying to do things. If we are [00:56] digitizing human experiences, really [00:57] what happens in our implementation is [01:00] that instead of our users needing to [01:02] have a team to speak to and review their [01:04] work and and collaborate with, we now [01:06] gave you the team. It started from this [01:09] mindset that okay, who is actually [01:10] sitting on the other side of these [01:12] conversations? It's an actual team [01:14] member, each one skilled in a specific [01:17] skill set. And that started leaning into [01:20] all right, let's just have a team of [01:22] team members. Each one is a one is a [01:23] designer, one is an SEO expert, one is [01:26] the project manager that kind of [01:28] wrangles everything in there, one's in [01:29] charge of voice, other in accessibility [01:32] and so on. With this mindset, then we [01:34] came into the challenge of how are we [01:36] going to present them. Is this just [01:37] going to be like a generic AI name and [01:39] we just say SEO, you know, or it could [01:42] have been exactly like that. uh which is [01:45] also fair you know and you see a lot of [01:46] these stuff in terms of how people [01:49] implement these types of AI agents where [01:51] it's just like a robot you know you [01:53] literally have an icon of a robot very [01:56] kind of generic now I also looked at it [01:58] from a perspective that our audience are [01:59] creative people for the most part they [02:02] want to have nice things they want to [02:04] have creative stuff surrounding them the [02:07] level of aesthetics that is required to [02:08] serve this audience is different [02:10] compared to other segments you know to [02:13] other types of people. Um, you can see [02:15] from my background that it's nice to [02:17] have a nice environment. You know, we [02:19] started thinking in the direction of all [02:20] right, there going to be characters. [02:22] Which characters? How are they going to [02:23] be structured? How is this going to [02:25] work? And really, the first version were [02:27] a few robots. This was the first [02:29] version. It was literally just like a [02:32] like a boring robot. We even had them as [02:34] characters here. A little bit more like [02:36] this, you know? And this is super [02:38] generic and kind of boring. Not creative [02:41] at all. It's just what you would expect. [02:42] Okay, it's AI. It's going to be a robot [02:44] that is smiling and maybe in different [02:46] colors or something like this. So, it's [02:48] very much like the safe approach, the [02:51] expected approach. But if you want to [02:53] have the mascot be memorable, it needs [02:55] to connect to something that is a little [02:56] bit beyond uh when it comes to the [02:58] company, the way that you're doing [03:00] things, there's a why that matters a [03:02] lot. And that's exactly what I tried to [03:04] play with this here. Said, "All right, [03:06] let's just try and start with the shape [03:07] and then just maybe the top bit here is [03:09] the nose." And so if that's the case, [03:12] what would the other parts be? But then [03:14] maybe there's still robots. So I kind of [03:16] threw in that little uh antenna here [03:18] that helps balance the head in terms of [03:22] this is the nose because this is kind of [03:24] like a counterwe if you think of the [03:26] base here, the top of the head. So that [03:28] was the idea and that creates that kind [03:31] of connection. If this is here, the nose [03:33] is here, then you you already have that [03:35] alignment of how to place the two items [03:38] to make this feel like a nose. Then I [03:40] just started playing with all kinds of [03:41] different concepts what it could be and [03:44] just threw a bunch of ideas onto the [03:46] page really really quickly just to to [03:48] experiment uh with all these kind of [03:49] stuff. Imagine this was pixel. [03:52] >> Yeah, like Kirby. Yeah, [03:54] >> it's hard to imagine now that this could [03:56] have been the characters. Uh but it's [03:58] funny, you know, even [03:58] >> it's funny because we can ask the AI [04:00] what we thought about their original [04:02] kind of concepts. [04:03] >> Now we can Yeah. At the time we didn't [04:04] have the AI to ask these questions, but [04:06] >> yeah. Is is that a baby baby glitch or [04:10] baby pixel? [04:11] >> Yeah, this was this was Glitch. This was [04:13] the first version of Glitch for sure. [04:15] And this was kind of claro and I tried [04:18] to look at them from a designer [04:19] perspective and you see this was Na'vi [04:21] but then he became a she and so it's [04:24] really an iterative process to try and [04:27] get it and you can see how the fidelity [04:29] and and the kind of ideas start to be [04:32] articulated as you as you scroll down. [04:34] You literally see the the thinking [04:36] process and how things shaped up. Yeah. [04:38] And this was the this was Glitch before [04:40] this. And I tried to give each one of [04:41] them characteristics and personalities [04:43] based on how they act and think. I would [04:46] even say beyond that that for most of [04:48] them I have a person in mind. And so [04:51] they're mocked after people that I know [04:53] that are actually in these roles and [04:55] they're a bit of characters as well. So [04:58] you just kind of create the caricature [05:00] of the person. For example, Lexi, she's [05:01] in charge of the content. So let's give [05:03] her this kind of words on her clothes. [05:07] So it connects with that mentality. Uh [05:10] if index is in charge of SEO, I wanted [05:12] to indicate speed. That's why that kind [05:14] of glasses and you have that kind of [05:16] analytics stuff on his shirt initially. [05:19] Clero is the project manager. He needs [05:21] to be the one that communicates with [05:22] clients as well. A little bit more [05:24] reserved, a little bit more professional [05:26] in that way. And of course, you have a [05:28] pixel here on the top that is the [05:31] frustrated designer. nothing is good, [05:33] you know, everything can always be [05:34] better and he can see what we're doing [05:37] wrong and that's why he's never [05:38] satisfied and has that kind of mentality [05:41] and also [05:41] >> sounds quite familiar, Vito. You know, I [05:44] was certainly pulling out certain [05:46] aspects of these characters that were [05:47] absolutely inspired by real life and [05:49] nationalities. Yeah, they're French. [05:51] That's great. [05:52] >> Yeah. Yeah. Exactly. But then I thought, [05:54] all right, these are going to be AI [05:55] characters, so maybe I can just leverage [05:58] AI to level them up. And this is exactly [06:02] what happened by bringing those basic 2D [06:05] characters into the game and letting AI [06:07] play with them. And I I was also [06:09] thinking about this from like as I was [06:11] prompting into different versions, let's [06:13] include some concepts of like [06:16] ruggedness. These characters has seen [06:19] things. They're not new, you know, and [06:21] so that's why you have all this grime [06:23] and dirt on the face because they [06:25] actually went through the struggles for [06:28] you, you know, before that. also [06:30] thinking about this from an art [06:31] perspective. How can I make them be art [06:34] pieces in their own right so got [06:36] influenced here from cubism h which [06:39] Picasso was always my favorite artist [06:41] even as I was a teenager. The first art [06:44] book that I got was a Picasso kind of [06:46] like art book that an architect uncle of [06:48] mine gave me when I was a very very [06:50] young later on when I was in high school [06:52] and was doing like plastic art as a [06:55] major. Picasso was a big influence on [06:57] how I was painting with oils and things [06:59] like this, you know, on canvases. And I [07:01] really like the concept of bringing the [07:03] entire perspective into the front so [07:06] that when you look at it from the front, [07:08] you might see the the ear that is on the [07:10] back and you might see the eye that you [07:12] can't see otherwise. And you know, so [07:13] bringing the whole thing into one place [07:16] and then you really don't know what's [07:17] behind. That's really is the one of the [07:19] core elements of cubism or specifically [07:21] Picasso's style in the later years at [07:23] least. And so yeah, so you can see all [07:25] of this stuff bundled into here.