#221 – Rahul Bansal on Using AI Everywhere at rtCamp ![rw-book-cover](https://wptavern.com/wp-content/uploads/2026/06/221-Rahul-Bansal-on-Using-AI-Everywhere-at-rtCamp.jpg) 🔗 **Source:** [#221 – Rahul Bansal on Using AI Everywhere at rtCamp](https://wptavern.com/podcast/221-rahul-bansal-on-using-ai-everywhere-at-rtcamp) > [!note]+ Document Notes > tags: [wordpress-agency, ai-automation, enterprise-wordpress, business-strategy, open-source-erp, specialisation] > author: Nathan Wrigley (host), Rahul Bansal (guest) > source: WP Tavern — Jukebox Podcast #221 > reference: https://wptavern.com/podcast/221-rahul-bansal-on-using-ai-everywhere-at-rtcamp > date created: 2026-06-17 13:30 > > # 🏷️ Title: #221 – Rahul Bansal on Using AI Everywhere at rtCamp > <!-- deep-read:sections start --> ## 🧭 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 - [rtCamp](https://rtcamp.com) - [Frappe / ERPNext](https://frappe.io) (open-source ERP rtCamp standardized on) - [WordCamp Asia 2026](https://asia.wordcamp.org) - [WordPress.tv](https://wordpress.tv) (talk video, when posted) ## 📝 Summary - **What happened.** Nathan Wrigley interviews Rahul Bansal, founder/CEO of rtCamp (17 years old, ~250 people, serves Google and Fortune 500s), expanding on his WordCamp Asia 2026 talk "How to Start an Enterprise WordPress Agency in 2026." The conversation moves from agency-building lessons (hire your complementary set; pick clients who need your expertise; niche within niches) to rtCamp's headline shift: an "AI everywhere" mandate — *if something can be done by AI, it will be.* - **Why it matters.** This is a concrete, founder-level account of restructuring an entire WordPress agency around AI — not a bolt-on feature but the operational backbone. The mechanism is specific: years ago rtCamp consolidated its back office onto open-source Frappe/ERPNext (one clean "single source of truth"), and that unified, first-party data is now the lever that lets them apply AI more aggressively than competitors stuck on scattered SaaS (Jira + QuickBooks + …). - **Key stakeholders.** WordPress agencies (especially enterprise/migration shops), agency founders starting in 2026, rtCamp staff facing role shifts, and the broader WordPress ecosystem positioned as an "AI winner" because open source exposes structured data without proprietary walls. - **Risks and tradeoffs.** Headcount addition has slowed (≈one-third of WordPress engineers currently under-utilized while work rises); coding-only roles are being devalued in favor of "growth engineers"/consultants/prompt-and-context engineers; next 50 hires are sales & marketing, not engineering. AI "hallucination" is framed as the volatile "jet fuel" cost that makes fixed-bid pricing hard. Rahul repeatedly credits luck, which limits the prescriptive value. - **Bottom line.** Don't invent AI — *apply* it. Take a "boring" high-volume problem (WordPress migrations), drive its cost down with AI on a unified data backbone, pass savings to clients, and the bigger pie (more migrations) more than offsets shrinking per-project hours. ## 📖 Main Content - **Hire your complementary set.** rtCamp's most-remembered idea. When moving from freelancer to agency, hire people who are the *opposite* of you, not copies. Rahul (strong engineering, weak sales/English) deliberately listed his weaknesses and hired against them; early hires were even asked questions whose answers he couldn't himself judge. Inspired by set theory from a childhood math class. Replication ("20 React engineers") comes only later, at scale. - **Complementary clients.** The same logic applies to client selection: sell to clients who *don't* understand WordPress, so your expertise stays premium and uncommoditized. Selling to agencies like yours crushes margins. - **Niche within niches.** rtCamp's original edge — "we only do WordPress" in a India full of generalist outsourcing — no longer works because "WordPress is the web now." New agencies must specialize far deeper: e-commerce → WooCommerce → payment-gateway/ERP/subscription specialization. Start with the smallest possible niche and grow bottoms-up. - **Why WordPress (and the role of luck).** rtCamp began as a media company / tech blog network (a "stripped-down TechCrunch") running on WordPress; Rahul refused to add Joomla (lost a co-founder over it) on the principle "we only sell what we use" (dogfooding) plus an open-source commitment. The blog's huge traffic forced them to write scalable WordPress code (chose Nginx early, scaled on a $10 Linode), which attracted their first customers — competitor tech bloggers. They drifted from blogging into custom WordPress development because it made more money. Rahul calls the WordPress bet "just luck" in hindsight. - **The pre-AI groundwork.** Post-COVID growth pushed rtCamp to "build the McDonald's of consulting" (inspired by *The Founder*) — codify and repeat every process. They consolidated accounting, payroll, project management, and CRM onto open-source **Frappe/ERPNext**, creating one unified, cleaned system of record. Intent then: scale to thousands of people. Then AI arrived and the headcount-scaling goal evaporated — but the unified first-party data turned out to be the real asset. - **The AI leverage thesis.** Because all rtCamp data (client, project, human, metadata) lives in one open-source system, it's "a sitting duck to query in any way," unconstrained by SaaS vendors' enterprise/AI tiers. That's their edge over agencies scattered across 6–7 tools. - **Where AI is applied.** Lead qualification (auto-research on form submit, replacing a 30-min junior task — that entry-level sales job is gone); proposals (vibe-coded live WordPress Playground demo sites instead of PPTs/Figma, even copying a client's design system for migrations); discovery/data-mapping in minutes; effort estimates blind-cut ~30% (worst case AI gains 20%, best case >90%); one migration that took 10 days now done in 5, including automated QA bots opening random Zendesk tickets to verify metadata/deploys. - **Pricing & the bigger pie.** Clients want fixed-output bids, not hourly. AI hallucination is the volatile "jet fuel" cost that complicates fixed pricing. Lower cost of sale → lower minimum quotes (e.g., AEM migration floor dropping from ~$100k toward ~$50k). Cheaper, faster migrations bring fence-sitters off the fence; demand rises. WordPress wins because open source keeps up with fast AI and exposes structured data without proprietary walls — they're seeing *more* migration inquiries amid the AI boom. - **Role and team reshaping.** "Coding is a race to the bottom." Value shifts to people who can *imagine* what to build and choose among 20 execution paths — prompt/context engineering. rtCamp's "Growth Engineers" (top coders) now go on first client calls to make feasible, non-over-promised commitments. Net headcount growth slowed (200→250, next 50 "very slow"); first time the careers site has no engineering openings; the next 50 hires will be sales & marketing, shifting the mix from ~10% to ~30%. They're explicitly choosing not to lay people off (unlike a shareholder-owned company that might "fire a hundred"), redeploying staff and inviting them to find parts of the business to optimize. - **The core philosophy.** "We are AI's consumers/consultants, not its inventors — our job is to *apply* AI, not invent it." The failure mode is vibe-coding products nobody buys, then blaming AI. Instead: take existing, boring problems (rtCamp = "the movers and packers of the internet," moving sites from Sitecore/AEM to WordPress) and make them dramatically cheaper/faster at the same quality. ### 🗒️ Notes - **Complementary-set hiring** is the audience's #1 takeaway and Rahul says it's timeless — still how he'd start any agency (WordPress or AI) today. List your weaknesses; hire their opposite. - **Complementary clients**: pick clients who need what you know; avoid selling to peers (margin death). - **Specialisation**: "niche as small as possible, then go bottoms-up." WooCommerce → payment gateways/ERPs/subscriptions. - **"We only sell what we use"** (dogfooding) is the throughline: it explains both the WordPress choice and the Frappe/ERPNext-into-AI consulting path. - **Luck, named honestly**: WordPress over Joomla "could have backfired … in hindsight it was just luck." Nathan praises the rare humility. - **Single source of truth** (Frappe/ERPNext) built for one reason (scale headcount) became valuable for another (AI leverage) — Jobs's "connect the dots looking backward." - **70–80% figure** clarified: refers both to share of operations streamlined and to cost/time reduction; non-build "meta cost" (proposals, marketing, case studies) already roughly halved. - **Concrete AI wins**: lead-qual automation killed an entry-level sales role; vibe-coded Playground demos replace slide decks; a 10-day migration done in 5 with automated QA bots. - **Pricing physics**: hourly billing dying; fixed-output bids need cost consistency; AI hallucination = the uncontrollable "jet fuel" variable; KPI is "worst-case AI gain" (now ~20%, target ~70% in two years). - **Open source as AI moat**: WordPress exposes structured data without proprietary walls; "only open source can keep up" with AI's pace → more migration inquiries. - **Workforce**: 200→250, next 50 sales/marketing; careers site has zero engineering openings; "growth engineers" make the first-call promises; ~1/3 of WP engineers currently under-utilized. - **Aside**: Rahul name-drops "Opus 4.8" as the kind of thing agencies *don't* need to worry about inventing — "let cloud engineers think of that." ## 📚 References - [rtCamp — company site](https://rtcamp.com) - [rtCamp blog post on AI ("AI everywhere")](https://rtcamp.com/blog/) (the post Nathan references; specific URL in show notes) - [Frappe / ERPNext](https://frappe.io) — open-source ERP, CRM, project management - [The Founder (2016 film)](https://www.imdb.com/title/tt4276820/) — "McDonald's of consulting" inspiration - [WP Tavern Jukebox episode index](https://wptavern.com/podcast) ## 🗣️ Paraphrase - rtCamp is a 17-year-old, ~250-person enterprise WordPress agency serving Google and Fortune 500s; this episode expands Rahul's WordCamp Asia 2026 talk. - The most durable lesson: when building an agency, hire people who are your opposite, covering each other's weaknesses; replicate skill sets only at scale. - The same complementary logic picks clients — sell to those who lack your expertise so you stay premium. - Differentiation today requires niching far below "WordPress" into specific sub-domains and becoming the recognized expert there. - rtCamp chose WordPress because they used it (dogfooding) and were committed to open source; Rahul frankly credits luck. - Post-COVID they unified their back office on open-source Frappe/ERPNext to scale headcount; that clean single source of truth unexpectedly became their AI advantage. - The mandate now is AI everywhere — applied to sales, proposals, project management, and PoC builds — to cut operational cost, not to invent novel products. - Cheaper, faster migrations expand demand; WordPress wins as an open, structured-data platform that keeps pace with AI. - Roles shift from coding to imagining and prompting; hiring tilts toward sales/marketing/consulting; rtCamp redeploys rather than lays off. ## 📰 Newsletter **rtCamp's bet: AI isn't a feature, it's the backbone** If you run (or dream of running) a WordPress agency, this WP Tavern Jukebox with rtCamp's Rahul Bansal is required listening. Two ideas travel well beyond rtCamp. First, *hire your complementary set* — staff your weaknesses, not your strengths (he even hired people whose answers he couldn't grade). Second — and this is the spicy one — *AI everywhere*: "if it can be done by AI, it will be." The clever bit isn't the slogan, it's the prerequisite. Years ago rtCamp put accounting, payroll, PM and CRM onto open-source **Frappe/ERPNext** — one clean source of truth. That unified, first-party data is exactly what lets them point AI at the whole business while competitors juggle Jira + QuickBooks + five other SaaS silos. The payoff: lead-qual automation that retired a junior role, vibe-coded WordPress Playground demos in proposals instead of slide decks, and a migration that went from 10 days to 5 with QA bots. His framing for the rest of us: *apply AI, don't invent it.* Take the boring, high-volume work — migrations, "movers and packers of the internet" — make it cheap and fast, and watch the pie grow. And note the WordPress-specific thesis: open source exposes structured data without proprietary walls, so "only open source can keep up" with AI's pace. Optimism with receipts. ## ✅ Action Items #### DevRel - Developer advocacy: turn "apply AI, don't invent it" + the Frappe single-source-of-truth pattern into a talk/demo for WordPress agencies. - Education: workshop on building vibe-coded WordPress Playground demos for sales proposals. - Developer experience: document a reference "AI-ready agency data backbone" using open-source tooling. - Developer marketing: clip the "complementary set" hiring idea as evergreen content. - Community management: host an agency-owner roundtable on AI role-shift anxiety. - Internal/external/cross-project: share migration-automation patterns (QA bots, data mapping) across the ecosystem. #### OSPO - Policy & Governance: contribution/disclosure norms for AI-assisted agency deliverables. - License Compliance: confirm provenance when AI copies client "design systems" during migrations. - Security Management: data-governance review of first-party data piped into AI (PII in CRM/ERP). - Software Inventory: catalog open-source ERP/CRM stacks as alternatives to SaaS lock-in. - Developer Support: consultation on Frappe/ERPNext + AI integration. - Community Engagement: fund case studies on open-source data backbones for agencies. - Strategic Alignment: track "cost-of-sale reduction" as an ecosystem KPI. #### Community - Maintainers: Not directly applicable to this source. - Contributors: "growth engineer" career-path narrative for newcomers worried about coding-only roles. - Sponsors: report migration-volume growth as an ecosystem-health signal. - Ecosystem: coordinate shared migration tooling (AEM/Sitecore → WordPress). #### WordPress - Core: monitor whether AI-driven migration demand stresses importer/REST tooling. - Core Performance: migration automation needs reliable perf baselines — publish them. - Core AI: align "apply not invent" guidance; structured-data exposure as an AI advantage to document. - **Core Program Team:** coordinate cross-team response to the agency-AI shift. - Mobile: rtCamp shipped its first iOS app last month — note cross-platform demand signal. - Playground: rtCamp uses Playground for live proposal demos — promote this as a sales use case and harden it. - Community: WordCamp session track on "AI for agencies." - Training: learning path for prompt/context engineering aimed at WordPress devs. - Marketing: amplify the "WordPress wins because open" narrative. - Support: anticipate migration-surge support load. - Design: document design-system-copying ethics/limits in AI migrations. - Documentation: examples-first guide to exposing WP structured data for AI. - Photos / TV / Openverse: Not applicable to this source. - Polyglots: glossary for "vibe coding," "growth engineer," "single source of truth." - Accessibility: ensure AI-generated migration output meets WCAG (see Accessibility Assessment). - Meta: consider .org signals for AI-assisted agency listings. - CLI: migration automation often leans on WP-CLI — document robust patterns. - Hosting: migration surge implies onboarding-at-scale playbooks. - Tide / Themes / Plugins: Not applicable to this source. #### The WPCC - Programs: fellowship on open-source operational backbones for small agencies. - Projects: documentation sprint on AI-applied migration workflows. - Funding experiments: microgrants for agencies adopting open-source ERP + AI. ### FAIR - Technical independence: rtCamp's first-party, open-source data store is a textbook FAIR-aligned independence play — document it. - Policies and covenants: transparency on AI use in client deliverables and data handling. - Productized governance artifacts: an "AI-ready agency" charter + data-governance scorecard. ## 🧪 Hypotheses - H1: Agencies with a unified open-source data backbone realize larger AI cost reductions than agencies on fragmented SaaS stacks. - H2: "Complementary set" hiring correlates with higher early-stage agency survival than same-skill hiring. - H3: Lowering migration price floors (e.g., $100k → $50k) increases total migration volume enough to grow revenue despite lower per-project hours (elastic demand). - H4: Deeper niche specialization predicts higher margins than generalist "WordPress agency" positioning in 2026. - H5: AI hallucination variance (worst-case 20% → best-case 90% gain) narrows over time, making fixed-output bids more reliable. - H6: Agencies that redeploy rather than lay off retain more institutional knowledge and adapt faster to the consulting-led model. ### 📊 Methodology - H1/H3/H5: longitudinal agency study tracking stack architecture, per-project hours, win rates, and price floors against realized AI gains; rtCamp's own ERPNext data is an ideal dataset. - H2/H4/H6: multi-agency survey + cohort analysis on founding-team composition, niche depth, and retention vs. margin/survival outcomes. - H3 specifically: price-elasticity analysis of migration inquiries before/after AI-driven price-floor cuts. ### ❓ Questions - Does the "open-source data moat" hold once SaaS vendors ship competitive AI query layers? - What's the floor on migration price before quality or sustainability breaks? - How transferable is rtCamp's playbook to agencies *without* a pre-existing unified data backbone? - Where does "apply not invent" stop being enough as AI commoditizes application too? ### 🔮 Future Research - Empirical price-elasticity of WordPress migration demand under AI-driven cost cuts. - Comparative study of open-source ERP vs. SaaS stacks as AI-readiness predictors. - The economics and ethics of AI copying client design systems during migration. - Workforce study on the coder → growth-engineer transition (reskilling success rates). ### 💡 Implications - **Practice:** Consolidate operational data *before* chasing AI; the backbone is the prerequisite, not the AI tool. - **Policy:** Disclosure and data-governance norms for AI in agency deliverables become competitive necessities. - **Security:** First-party data piped into AI raises PII/governance stakes. - **Education:** Curricula must pivot from coding to prompt/context engineering and solution imagination. - **Funding models:** Fixed-output bidding + AI-variance KPIs reshape how agency work is priced and funded. ### 🧠 Learning Styles Assessment - Reading level estimate: conversational spoken transcript; ~Flesch–Kincaid Grade 8–10 (long, winding sentences but common vocabulary; business jargon: "single source of truth," "vibe code," "complementary set"). Flesch Reading Ease ~55–65. #### 👀 Visual learners - Present: none — audio-only podcast transcript, no diagrams. - Improve: a diagram of the unified-data-backbone → AI-leverage flow; a before/after migration-timeline graphic; the hiring "set theory" Venn diagram. #### 🔊 Aural learners - Present: strong — it's a podcast with clear narration and a host summary. - Improve: chapter markers and a 3-minute audio recap. #### 🗣️ Verbal learners - Present: vivid analogies ("movers and packers of the internet," "McDonald's of consulting," "jet fuel" hallucination, set theory). - Improve: a glossary callout (vibe coding, growth engineer, single source of truth, fixed-output bid). #### 🧩 Kinesthetic learners - Present: none hands-on. - Improve: a worksheet to map your own "complementary set" weaknesses-to-hires; a Playground demo-build exercise. #### 🧠 Logical learners - Present: clear cause→effect (unified data → AI leverage → lower cost → bigger pie). - Improve: a decision tree for niche selection; a pricing model showing elastic-demand math. #### 🧑‍🤝‍🧑 Social learners - Present: interview/dialogue format. - Improve: discussion prompts for agency-owner peer groups. #### 🧘‍♀️ Solitary learners - Present: reflective founder narrative. - Improve: journaling prompts ("what boring problem could I make 50% cheaper?"). #### 🌿 Naturalistic learners - Present: "rising tide / bigger pie" ecosystem framing. - Improve: tie open-source/first-party data to data-minimization and sustainability of compute. ### 📊 Accessibility Assessment #### Screen reader semantics - Present: WP Tavern page uses headings and a structured transcript. - Improve: ensure speaker labels are marked up semantically; descriptive headings per topic. #### Keyboard navigation - Present: standard web page; audio player should be keyboard-operable. - Improve: verify player controls have visible focus and labels. #### Contrast and readability - Present: WP Tavern's theme generally meets contrast norms. - Improve: verify 4.5:1 on player UI and link colors. #### Text resizing and reflow - Present: responsive transcript text. - Improve: confirm no clipping at 200% zoom on the embedded player. #### Media and captions - Present: a full text transcript accompanies the audio — strong (WCAG 1.2.1). - Improve: add timestamped chapter navigation; downloadable transcript. #### Forms and interactions - Present: contact/subscribe forms referenced. - Improve: ensure form labels and inline validation. #### Standards reference - Maps to WCAG 2.2 AA: 1.2.1 (transcript present ✓), 1.4.3 Contrast, 2.1.1 Keyboard, 1.4.10 Reflow, 2.4.6 Headings/labels. ### 🌍 DEIB Assessment #### Cultural breadth - Present: strong global lens — India-based agency, references to outsourcing history, global enterprise clients; a candid "(might sound inhuman)" self-check on the Dubai-travel analogy. - Improve: more non-founder voices; perspectives from affected staff. #### Inclusive language - Present: generally inclusive, plain; jargon mostly explained. - Improve: define "vibe code," "growth engineer," "single source of truth" inline; soften the layoffs-as-virtue framing for affected workers. #### Neurodiversity support - Present: long monologues, few signposts. - Improve: add subheadings/timestamps and chunked summaries for the long answers. #### Representation - Present: two voices (host + founder), both senior men. - Improve: include junior-staff or differently-positioned agency perspectives. #### Psychological safety - Present: acknowledges that change "is scary for us also," commits to not laying off, communicates openly. - Improve: address worker anxiety directly with concrete reskilling support details. ### ✅ Logical Validity #### Strong reasoning - Causal (unified data → AI leverage): well-mechanized — single, queryable, first-party data store explains the competitive edge cleanly. - Analogical: "movers and packers of the internet," "McDonald's of consulting," and the "jet fuel = hallucination" analogy are apt and clarifying. - Inductive (elastic demand): lower migration price → more migrations is plausible and partly evidenced ("more migration inquiries with the AI boom"), though not quantified. - "Connect the dots backward" framing: honest acknowledgment that the data backbone's value was realized only retrospectively. #### Fallacies - **Appeal to luck / survivorship bias** — repeated "it was just luck" about the WordPress bet. The narrative is told by a survivor; agencies that made similar bets and failed aren't represented. Fix: bound lessons to what's repeatable (dogfooding discipline) vs. what's contingent (timing/luck). - **Hasty generalization** — "70 to 80%" of operations streamlable, and "build time will fall by 70% in two years," generalized largely from rtCamp's own (atypically prepared) situation. Fix: present as rtCamp-specific projections with stated assumptions. - **False analogy (acknowledged)** — the Dubai-travel elasticity analogy; Rahul himself flags it as "a bad example, might sound inhuman." Fix: replace with a like-for-like B2B example. - **Optimism / motivated reasoning** — "WordPress will emerge as a winner … only open source can keep up with AI." A strong, agency-interested prediction stated with high confidence. Fix: frame as a hypothesis with falsifiers (e.g., SaaS vendors shipping competitive AI query layers). #### Tone of voice - Reflective, humble (frequent luck attributions), optimistic-but-realist, founder-practical; Nathan's prompts are warm and probing. Mild self-promotion balanced by candor about staffing friction. ### 🛡️ DARVO Analysis - Deny: Mild — the layoffs reality is softened ("we don't want to fire people," redeploying instead), but it is acknowledged rather than denied (one-third of WP engineers under-utilized is stated plainly). - Attack: None — no accusers; he gently critiques peers who "vibe code things they can't sell," but as caution, not attack. - Reverse victim/offender: None. - Overall: DARVO is largely not applicable. The closest tension is reframing workforce disruption as optimism/opportunity, but Rahul names the discomfort openly ("change is scary for us also"), which weighs against a defensive DARVO reading. ## 🎯 Conclusion - **Bottom line.** rtCamp's edge isn't "using AI" — it's having consolidated its operations onto one clean, open-source, first-party data backbone *before* AI, then applying AI across that backbone to slash cost-of-sale and build cost. The strategy: apply AI to boring, high-volume work (migrations), pass savings on, and grow the pie. Roles shift from coding to imagining/prompting; hiring tilts to sales/consulting. - **Next steps for readers.** (1) Audit how scattered your operational data is — consolidation is the AI prerequisite. (2) Identify one "boring" high-volume service you could make ~50% cheaper with AI. (3) Pilot vibe-coded Playground demos in proposals. (4) Reframe roles around solution-imagination and prompt/context engineering, and plan redeployment, not just hiring. ## 🔄 Refactor Appendix — Rahul Bansal on Using AI Everywhere at rtCamp > Density: 2–5 Literature atoms (H3), 0–3 Permanent candidates (H4). > Defaults applied. Resonance filter on. > > Tag pool: #bmf · #addendum · #xref · #articleseed · #storystem · #share · #aphorisms ### 📚 Literature Atoms (H3) #### [H1] Hire your complementary set, not copies of yourself > [!quote] ID:rw_pending > "I literally listed down my weakness and found people who were opposite of me." **Context & use**: - Audience's #1 takeaway; "timeless," still how he'd start any agency today. - Replicate skill sets only later, at scale. **Tags**: #xref #aphorisms #share **Highlight**: [view highlight](readwise://pending) <!-- atomicity: 1 idea ✓ --> #### [H2] A unified open-source data backbone is the AI prerequisite > [!quote] ID:rw_pending > "The data is first party. Like sitting duck there to query in any way we can. We are not limited by SaaS providers, enterprise plan or this AI capability." **Context & use**: - Frappe/ERPNext single source of truth, built pre-AI to scale headcount, became the AI lever. - Edge over agencies scattered across 6–7 SaaS tools. **Tags**: #xref #articleseed **Highlight**: [view highlight](readwise://pending) <!-- atomicity: 1 idea ✓ --> #### [H3] Apply AI to boring high-volume work to grow the pie > [!quote] ID:rw_pending > "We are agency. Our job is to apply AI, not invent AI." **Context & use**: - Make migrations cheaper/faster; lower price floors bring fence-sitters → more demand. - Failure mode: vibe-coding products nobody buys, then blaming AI. **Tags**: #xref #aphorisms **Highlight**: [view highlight](readwise://pending) <!-- atomicity: 1 idea ✓ --> #### [H4] Coding is commoditizing; value shifts to imagining what to build > [!quote] ID:rw_pending > "coding is race to bottom … You will need people who can imagine what needs to be built." **Context & use**: - "Growth engineers" make first-call promises; next 50 hires are sales/marketing. - Prompt/context engineering is the rising skill. **Tags**: #xref #addendum **Highlight**: [view highlight](readwise://pending) <!-- atomicity: 1 idea ✓ --> ### 🧠 Permanent Candidates (H4) #### [P1] AI advantage is downstream of data consolidation, not AI adoption **Claim** (one sentence, Courtney's voice): An organization's ability to extract value from AI is determined by how unified and first-party its operational data already is, so consolidating the data backbone is the real precondition for AI leverage. **Evidence pointer**: [[#H2]] **Scope & conditions**: Holds for process-heavy service businesses with structured operational data. Breaks where the core value is novel invention rather than repeatable process, or where data can't be unified for legal/privacy reasons. **Links**: - [[Single source of truth]] — because extends - [[SaaS vendor lock-in]] — because contrasts on data control - [[WordPress as open structured data]] — because applies-in the WordPress AI thesis <!-- atomicity: 1 idea ✓ --> ### ✏️ Fleeting Queue - [F1] Try a vibe-coded WordPress Playground demo as a proposal artifact for one real pitch — https://wptavern.com/podcast/221-rahul-bansal-on-using-ai-everywhere-at-rtcamp #fleeting #inbox - [F2] Open question: at what migration price floor does quality/sustainability break? #fleeting #inbox ### → Things 3 (max 2 actions) - [ ] Draft a short "apply AI, don't invent it — consolidate data first" note for the WordPress agency audience — #things - [ ] Add the "complementary set" hiring frame to the agency-advice resource — #things <!-- deep-read:sections end --> 📄 **Full text:** ![[Full Document Contents/Articles/#221 – Rahul Bansal on Using AI Everywhere at rtCamp.md|#221 – Rahul Bansal on Using AI Everywhere at rtCamp by wptavern.com]] ## Highlights - Building a WordPress agency business for large enterprises. In this talk, I’ll share the story of how rtCamp grew from a small WordPress shop into a globally recognised enterprise agency, trusted by Fortune 500 companies and major publishers. If you’re starting an agency today or looking to move up market in 2026, this session will give you a realistic roadmap building on real lessons from my personal experience. ([View Highlight](https://read.readwise.io/read/01kvbaz09rn2wvz404rpg94s31)) - Tags: 👻 ai highlighted - Rather than making it as a nice story about what worked for us, I lean more towards practical advice, and that’s where the essence remained. But I focus more on the 2026 part, because when we started, it was 2006. The first time when I used WordPress was 2006. rtCamp started in 2009. 20 years is a big time. And then at the end of this 20th year, like we are going through this AI led change. ([View Highlight](https://read.readwise.io/read/01kvbbkqzx45y7qqewn5s0d1m0)) - Tags: 👻 ai highlighted - So I literally listed down my weakness and found people who were opposite of me. Even interesting part was that, to the few initial hires I asked the questions, whose answer I had no idea whether they’re saying right or wrong. ([View Highlight](https://read.readwise.io/read/01kvba2vptsz86ba18tzndnf6v)) - Note: H1 — Hire your complementary set: list your weaknesses, hire their opposite. Audience's #1 takeaway; timeless. - Tags: hiring, agency-strategy, atom/literature/h1 - So that is why we can leverage AI more than a company, agency to agency. For agencies using say, Jira for project management. QuickBook for accounting, some other software. If their operations is scattered across 6, 7 software, we have leverage over them. Not only we are paying very less because all our software is open source. The data is first party. Like sitting duck there to query in any way we can. We are not limited by SaaS providers, enterprise plan or this AI capability. ([View Highlight](https://read.readwise.io/read/01kvba2yppa5xnh44xqdpyndr6)) - Note: Strongest atom + Permanent P1: unified open-source first-party data is the AI prerequisite/moat vs. agencies scattered across 6-7 SaaS tools. - Tags: atom/zettel/z1, atom/literature/h1, single-source-of-truth - Like, what people are doing is everybody’s trying to act like a ChatGPT, OpenAI, it’s their job to invent AI algorithm. We are agency. Our job is to apply AI, not invent AI. We don’t have to think of what is Opus 4.8 will do. Let cloud engineers think of that. ([View Highlight](https://read.readwise.io/read/01kvba30t3ct34g4ecvx7syvvs)) - Note: H3 — "Our job is to apply AI, not invent AI." Failure mode: vibe-coding products nobody buys, then blaming AI. - Tags: aphorism, atom/literature/h2, ai-strategy - That is what our internship look like, because coding is race to bottom. Eventually the cost of building will shrink to the point that you don’t need many, you won’t need many traditional developers in any agency. You will need people who can imagine what needs to be built. There might be 20 different ways and which way this project should be executed. That prompt engineering, context in engineering. ([View Highlight](https://read.readwise.io/read/01kvba3590me8avt01r5mex8aj)) - Note: H4 — Coding commoditizing; value shifts to imagining what to build + prompt/context engineering. Drives rtCamp's sales/marketing hiring tilt. - Tags: prompt-engineering, future-of-work, atom/literature/h2 ## Links - [[AI's Leverage Comes From the Data Foundation Not the Model]] — existing Permanent captures rtCamp's core advantage from consolidating first-party operational data before adopting AI. - [[AI Erodes the WordPress Plugin Value Props Pushing Monetization Toward SaaS and Services]] — applies-in: cheaper coding moves agency value toward judgment, consulting, and business outcomes.