Agenteers in the Wild
Twenty real Agenteers and the agents they've built — across finance, marketing, ops, healthcare, law. The companion field guide to 'What is an Agenteer?'.
This is the companion piece to "What is an Agenteer?". The essay defines the role. This article shows the people.
Twenty-plus named Agenteers, profiled with the four things every other Agenteer wants to know: how technical they are, why they built what they built, what they actually built, and which tools they used. The list will keep growing. If you're building, send your story.
In the Big Four & professional services
Anri Coetzee — Director, Global Process Owner (General Ledger), EY
Anri is a Chartered Accountant (CA(SA)) running global general ledger processes at EY — a Big Four firm with nearly 400,000 employees. Her background is finance, not engineering: she's spent her career inside accounting operations.
The why: Journal entries at EY were managed differently in every country, with country teams emailing spreadsheets to global processing teams. Lead times stretched into days. The work was unloved and unscalable.
What she built: PowerPost — a unified journal entry automation that replaced the email-and-spreadsheet chain with a single self-service workflow. Country teams submit through the app; validation, routing, and posting happen automatically. Paula Korczak (Product Incubation Manager) later extended the system with a Copilot Studio AI agent — turning what used to be a 15-minute submission for 20 journals into something closer to a minute.
The stack: Microsoft Power Platform (Power Apps + Power Automate). Copilot Studio for the agent layer.
The numbers: 95% reduction in GL lead times. $4M+ projected savings over five years. 2,000+ EY employees using the tool globally. Won SAICA's Chairman's Difference Makers Awards (Innovator category, 2022).
"PowerPost has fundamentally reshaped how we manage journals across EY."
(The Microsoft case study credits the quote to "Anri Brits" — her maiden name. Same person.)
Marie Sanders — Global Process Owner (Accounts Receivable), EY
Same firm, different domain. Sanders is a finance operations process owner. She built a sister tool to PowerPost focused on AR.
The why: Payment matching at EY was painfully manual. Customer payments came in via wire, ACH, and check; invoices sat in NetSuite waiting to be matched; humans cross-referenced everything. The rate of automated matching hovered around 30%.
What she built: PowerMatch — a payment clearing and cash application engine that ingests incoming payments, matches them to open invoices, and posts the cash automatically. Exception handling routes only the genuinely ambiguous cases to humans.
The stack: Microsoft Power Platform.
The numbers: Built in under four months. Payment matching went from 30% to 80%. 120,000 hours saved per year at 50% rollout — projected 230,000 hours at full global deployment.
"The fact that we were able to build the PowerMatch app in less than four months is amazing and really due to the low-code development platform we used."
Jamie Tso — Senior Associate, Clifford Chance (Hong Kong)
Jamie is a Senior Associate at Clifford Chance — one of the world's largest law firms (Magic Circle). He's not formally an engineer, but he taught himself programming on weekends after experimenting with TensorFlow. Most of his daily life is law, not code.
The why: Big Law work involves enormous amounts of repetitive document processing — redlining contracts, building signature packets, mapping fund prospectuses for due diligence. Each is high-stakes, but the underlying mechanics are pattern-matching. Tso started building tools for himself, then for colleagues who kept asking how he was getting work out the door faster.
What he built:
- RedlineNow — instant contract redlining (built in 2 days, December 2025).
- SignaturePacketIDE — automated M&A signature packet generation. Real usage in BigLaw deal teams.
- Fund prospectus mapping tool — his document-processing tools "went viral internally and were ultimately adopted by the firm."
The stack: Microsoft Copilot Studio, Power Automate, vibe coding with open-source AI tools.
The angle: By asking the AI to build a deterministic tool rather than answer a question inside the workflow, he reduces error rates compared to using AI live. The tool is reliable; the AI in the build process is the variable he's controlling.
"I got obsessed and started building not just for myself but for colleagues."
Read the Artificial Lawyer interview
In traditional industries
Tyler Diogo — Operations Manager, Arden Insurance Services
Tyler is the Operations Manager at Arden Insurance Services — a traditional insurance MGA (managing general agency). His background is operations and accounting; he found his way into automation entirely by accident, while running a Discord community for a video game on weekends. He needed Zapier to send Discord notifications and update calendars.
The why: Arden was sending tens of thousands of past-due invoice notifications per year by hand. Customer service reps were chasing customers for payment. Legal mailings were a manual queue. Tyler realized the same Zapier patterns he was using for his hobby community could rebuild the entire AR workflow.
What he built: A 12-Zap workflow chain for past-due invoice notifications and legal mailing. Once that worked, automation became the company's operating philosophy. Tyler now leads Arden's broader automation program — touching billing, customer communications, internal reporting, and compliance.
The stack: Zapier (with AI logic embedded in some workflows).
The numbers: 17,000+ hours of work automated in 2023, projected to double to 34,000 in 2024. $500,000+ in annual overhead savings. $150 million in annual company billings flow through automated processes.
The cultural shift is the most quotable result. As his customer service reps reported in company-wide meetings:
"Our customers now chase after us to pay their invoices, rather than us chasing after them."
Read the Arden Insurance story
David Gabriel — VP Marketing, Rhumbix
Rhumbix is a 12-year-old construction technology company recently acquired by Autodesk. David is VP of Marketing — a career marketer, not an engineer. Rhumbix went through a deliberate AI-native transformation under CEO Zach Scheel, with help from an outside AI advisor. David's sales-team build is one piece of that broader cultural shift.
The why: Sales reps were juggling three tools (n8n, Make, Zapier) plus mobile workflows for prospect research, meeting prep, and follow-up. The cost was rising and the experience was fragmented. David wanted one system that worked end-to-end.
What he built: A Lindy AI Assistant deployed across the entire Rhumbix sales team — handling email triage, meeting preparation, proactive scheduling, mobile workflows via text, deal tracking, and meeting-notes review. Five-minute setup per user. Zero engineering involvement.
The stack: Lindy as the orchestration layer, connecting to Salesforce, Gmail, calendar, and the broader sales stack natively.
The numbers: $25,000/month in cost savings vs. manual workflows. 80% of go-to-market work delegated to AI agents. Three previous tools consolidated. Full sales team onboarded within days.
"Deploying Lindy felt like flipping a switch. I don't know any AI agent that has taken me five minutes to build, even using Claude Code."
Listen to the full Rhumbix podcast | Lindy case study
Caren Kelleher — Founder, Gold Rush Vinyl
Caren is the founder of Gold Rush Vinyl — a vinyl record pressing plant in Austin, Texas, running 24-hour production with a small team (1-50 employees, per Zapier). Her background is partnerships at Google (Head of Music App Partnerships) and a Harvard MBA. Zero coding background.
The why: A vinyl pressing plant has a chaotic operational layer: customer order forms, file delivery, production scheduling, customer status updates, shipping. Without automation, Gold Rush couldn't have grown. Caren needed her tiny team to operate like a much larger one without expensive ops hires.
What she built: Over time, 76 active Zaps automating 18,000+ tasks per year — order intake, customer onboarding, file QA, production handoff, status notifications, shipping coordination.
The stack: Zapier, with integrations across Google Workspace, Stripe, customer support tools, and the production system.
The numbers: 2,285 hours of manual work saved per year (≈1.2 years of labor). 50% leaner headcount than the business would otherwise require.
"We couldn't do our work without Zapier. The work [Zapier] is doing helps us do more work for musicians."
Read the Gold Rush Vinyl story
Markus Kronen — Head of GenAI in Purchasing and Supplier Network, BMW Group
BMW Group has roughly 155,000 employees. Kronen started his career as a product owner working on AI quality before moving into GenAI leadership for the procurement organization. He sits at the boundary of technical and domain — closer to a power user than a software engineer, but with enough technical fluency to lead a multi-agent system program.
The why: Procurement at BMW means thousands of supplier evaluations per year, with quality data, pricing, supplier behavior, and process compliance to manage. Procurement specialists were swimming in data sources. The team needed an interface that let them ask questions in natural language and get reasoned answers across the full data fabric.
What he built: AIconic — a multi-agent system spanning 10 specialized agents covering quality, purchasing and supplier data, and purchasing process support. Procurement specialists query the system in natural language and the right agent handles the request, sometimes orchestrating across multiple sub-agents.
The stack: AWS Bedrock with Claude Sonnet (orchestration/reasoning) and Claude Haiku (classification).
The numbers: 1,800+ active users across BMW's procurement organization.
"Our multi-agent system AIconic significantly increases employee efficiency and productivity while setting new standards for AI usage."
In tech & startups
Ondrej Machart — Head of Product, Livesport (Flashscore)
Livesport is the Czech media company behind Flashscore — a sports data platform with 155+ million monthly users. Ondrej runs core product. His background is product, not engineering — including prior UX design experience. He started using Claude Code in September 2025 and has been documenting his journey publicly on Medium ever since.
The why: Like every product head, Ondrej had a backlog of internal tools that would help his team but never made it onto a roadmap. Engineering capacity was reserved for the public-facing product. With Claude Code, he stopped waiting.
What he built (highlights from 13 projects in six months):
- Product Portfolio Coach — AI-powered similarity analysis that helped inform a C-level brand consolidation decision.
- iOS prototyping baseline — reusable starter codebase that cut UI-cloning time across the team.
- Editorial infographics tool — close to production at time of writing.
- Data presentations system — Claude-generated web presentations with Confluence export. Used for major decisions including a Huawei device support sunset.
- Vibe-coding workshops — taught colleagues how to ship their own internal tools. Result: a team capacity dashboard, a design-system monitor, and others — all built by non-engineers.
The stack: Claude Code as primary. Hands off to senior engineers when something needs to move into production with formal QA.
The framing: This isn't magic, it's work. AI gets you far, but it isn't free, effortless, or risk-free.
Read his full 13-project breakdown
Hiba Fathima — Marketing Specialist, Firecrawl
Hiba runs marketing at Firecrawl, an AI tooling startup. Her formal title is Marketing Specialist; her actual scope is full-stack marketing — SEO, content, brand. She has a CS degree, which she says made the Claude Code learning curve easier — but most of her work was previously gated by designer and developer time.
The why: Small marketing teams at fast-moving startups have a familiar pain: every campaign or landing page requires designer time, developer time, and a queue of dependencies. Hiba wanted to remove the queue.
What she built:
- Autonomous brand-mention tracker — runs every Monday. Scrapes Reddit, Hacker News, Twitter, and niche forums for Firecrawl mentions. Surfaces relevant discussions for the week's content planning.
- Direct landing-page shipping — bypasses the designer/developer queue entirely. She drafts copy, structure, and assets; Claude Code builds the page; she ships it.
The stack: Claude Code as the daily driver. Firecrawl's own scraping API for the brand-mention work.
"If I know exactly what needs to go on a page, I can just... ship it. No waiting on anyone."
Read Firecrawl's full breakdown
Aditya Vempaty — VP Marketing, MoEngage
MoEngage is a Series F customer engagement platform (~$100M ARR). Vempaty leads marketing for North America. He's a senior marketer, not a developer — but he's spent enough time with Claude Code that he's now skeptical of the broader hype around AI marketing tools.
The why: AI-generated marketing copy has a specific problem: it sounds like AI. Anyone reading enough drafts can spot the patterns — the rhythms, the safe phrasings, the structural tics. Vempaty wanted a way to systematically catch those patterns and rewrite them.
What he built: A /humanizer skill in Claude Code that scores marketing copy drafts for "AI-ness" across four scoring categories, flags the offending patterns, and rewrites them in his voice.
The stack: Claude Code (skills feature).
The framing: Vempaty's value comes from a candid take that cuts against the broader marketing of these tools. He's tested most of them and concluded that some are doing real work and others are vapor.
"Claude Code does a job. Chat does a job. I honestly don't understand what Cowork is for."
Adam Schoenfeld — CMO, Inflection
Adam was CEO and co-founder of Keyplay until Inflection acquired the company. He's now CMO at Inflection and decided to skip the standard 30/60/90 onboarding plan, building an AI-native CMO operating system from day one — not as theory, but as the actual way he runs his job.
The why: Most new executives spend months reading internal docs, building mental models, and slowly assembling working knowledge of the company. Adam decided that work should live in machine-readable form from the start, where any agent (and any future hire) could draw from it.
What he built:
- Claude Code personal operating system — three-bucket local file architecture:
context(positioning, ICP, voice),work(active projects),system(the skills he reuses). /morningskill — pulls a personalized briefing from Slack, Notion, Calendar, recent customer calls, Salesforce, and LinkedIn.- GTM Context knowledge base — centralized, structured docs (positioning, customer stories, ICP definitions, win/loss analysis) that other skills draw from.
/draft-contentand/iterate-contentskills — with curated voice files for different audiences.- A new company website at inflection.io, built alongside everything else.
The stack: Claude Code as the operating layer. Slack, Notion, Salesforce, Calendar, LinkedIn as data sources.
The framing: Schoenfeld's biggest insight is that AI agents are only as good as the context they can pull from. The boring work of organizing context is the high-leverage work.
"AI agents are only as good as the context they have. The boring work of organizing and maintaining context is one of the highest-leverage things I've invested in."
Sebastian Volkis — Indie founder
Sebastian studied physics and astrophysics at university, became a freelance media buyer, then frustrated by the gap between "I have an idea" and "the engineers built it," he started using AI coding tools. He now ships SaaS products entirely with AI-coded code.
The why: Sebastian's media-buyer brain saw demand patterns engineers wouldn't necessarily prioritize. Content creators needed a way to find trending viral articles. SMBs needed AI customer support without an enterprise procurement process. He could see the opportunities; what he lacked was the artifact.
What he built (selected):
- TrendFeed — aggregates viral news, generates short-form content via AI. Built end-to-end in 4 days with Cursor.
- ChatIQ.ai — AI customer support chatbot. Originally built on Bubble (no-code) in 2023; later iterations using Cursor. 11,000 users, £7.2K MRR.
- CodeSpring — his AI coding course, £45K in revenue since January 2025.
The stack: Cursor as primary builder for new SaaS products. GPT-4 and Claude inside the apps. Bubble for the original ChatIQ.
The numbers: £9,222 (~$12K USD) in TrendFeed's first four weeks. £5.5K on launch day alone. He documented the journey publicly on Indie Hackers — including a public bank account screenshot to prove launch-day revenue.
"I built this entire app using AI in just 4 days."
Sabrine Matos — Founder, Plinq (Brazil)
Sabrine spent her career in growth marketing in Brazil, embedded in product teams but always on the sideline of engineering. She's not technical. Plinq is her first company.
The why: A woman in Sabrine's social circle was murdered by a partner with a hidden history of violent criminal convictions. In Brazil, women had no consumer-accessible way to check criminal records before dating someone. The market gap was lethal.
What she built: Plinq — a women's safety app that runs instant background checks on individuals. Risk scoring (green/yellow/red flags), panic button, and safety content feed. A B2B module for HR departments was planned for a later launch. Built end-to-end on Lovable in 45 days — frontend, backend workflows, integrations with public records databases.
The stack: Lovable for the entire stack. No supplementary frameworks, no external development.
Don't sleep on the next one. 2,480 Agenteers see it before it's public. Unsubscribe in two clicks.
The numbers: 10,000+ users within 3 months of launch. R$2.2M (~$456,000) in ARR. 300% month-over-month growth. 200+ dangerous situations identified and potentially prevented. Raising R$1.5–2M pre-seed with angel investors.
"If Lovable didn't exist, Plinq would never have seen the light of day."
Nomiki Petrolla — Founder, Theanna
Nomiki spent 15 years in product leadership, most recently as Head of Product at an AI startup. She's a non-technical product person — she lived through the engineering bottleneck for over a decade. Theanna is her solo company.
The why: Women founders face a network and access gap that men don't. Nomiki wanted to build a single platform — guidance, community, revenue tracking, a build interface — and she was tired of waiting weeks for engineers to ship features she could fully spec herself.
What she built: Theanna — an all-in-one platform helping women tech founders go from idea to $1M ARR. AI-powered guidance, revenue tracking, community, challenge prompts, and an in-app build interface. She personally built the entire app frontend using Claude Code in her terminal — natural language prompts and screenshots — handing off to engineers only for backend connections. The marketing website (with 12 third-party integrations) was built in Lovable in a single day.
The stack: Claude Code for the production app frontend. Lovable for the marketing website. GitHub for version control. Linear for project tracking. She has a public Lovable-vs-Claude-Code comparison breaking down when to use each.
The numbers: $207,506 ARR as of March 2026. 300+ women founders on the platform. Development cycle reduced ~90% (2 weeks to under 2 days) for shipping new features.
"Every single screen, every flow, every interface that the founders on Theanna interact with was built by me."
In healthcare & nonprofit
Arun Nadarasa — PCN Clinical Pharmacist, NHS (UK)
Arun is a working clinical pharmacist within the UK's National Health Service — an organization with roughly 1.5 million staff. His training is clinical, not technical. He spent years getting stuck at NHS hackathons because he couldn't code; he'd team up with developers to brainstorm healthcare solutions but couldn't build anything himself.
The why: NHS clinicians have ideas for tools that would make their work better — patient triage navigators, blood-test interpretation assistants, GP practice performance trackers — but they don't have engineers and they can't build the tools themselves. The gap between clinical insight and software was nearly absolute.
What he built (since discovering Lovable in September 2024):
- 20+ healthcare apps built without coding — a QOF Progress Tracker (GP practice performance monitoring), a Smart Blood Test Assistant, a Patient Triage Navigator, and others.
- Clinical AI Hackathons — he organized hackathons across NHS sites that drew 300+ registrants across 7 UK locations. 80 sign-ups in the first 6 hours of one event.
The stack: Lovable for everything. He builds with conversational prompts ("connect to this external API," "add a card to display patient information") rather than touching code directly.
The framing: Arun went from clinical-attendee-of-hackathons to organizer-of-hackathons in less than a year. He's become an internal NHS evangelist for AI building among clinicians. His own description of the original obstacle: he had no coding skills.
Colin Budries — Head of Support, Truemed
Colin runs customer support at Truemed — a specialized payment processor for HSA/FSA health and wellness purchases (~50 staff). His background is military (US Army Lieutenant) and accounting (Baylor University). He's not technical in the engineering sense, but he's analytical and structured by training.
The why: Support tickets at Truemed cluster around HSA/FSA eligibility questions, compliance queries, and standard account issues. Most have well-defined answers. The team didn't need more support reps — they needed an AI layer that could resolve the predictable cases and route the unpredictable ones to humans.
What he built: Customer-support AI agents that classify, route, and auto-resolve health-related tickets. Compliance-aware (HSA/FSA rules are intricate). Human escalation for ambiguous or unusual cases.
The stack: Lindy as the agent framework. Connects to Truemed's support system, knowledge base, and customer records.
The numbers: 5,000+ tickets automated. 36% of total support volume. 67% cost reduction per ticket ($1.00 → $0.33). Zero engineers needed for implementation.
"Lindy is like a developer that works just for me."
Justin Steele — Founder, Kindora
Justin spent ten years as Director of Google.org Americas, where he oversaw nearly $700 million in philanthropic giving. He has an MBA and MPA dual degree from Harvard Business School and Harvard Kennedy School, plus a chemical engineering undergrad from UVA — but he had never built web applications before starting to code with AI tools. After leaving Google.org, he founded his own nonprofit (Outdoorithm Collective) and ran into a wall.
The why: Existing grant-matching platforms returned thousands of poorly matched funders, leaving program officers to do the triage by hand. Justin had spent ten years on the funder side, watching how grant decisions actually get made. He knew exactly which signals separated good matches from noise — and he knew the existing tools weren't using those signals. The expertise was his moat. He just needed to ship the artifact.
What he built: Kindora — an AI-powered fundraising platform for nonprofits. Funder match scoring, grant proposal writing assistance, funder research briefs, voice-based pitch practice. The matching engine filters 3,000 raw matches down to ~75 viable prospects per nonprofit (a ~97.5% reduction).
The stack: Claude Code for the build itself. Claude Sonnet 4.6 (assistant, outreach, research), Haiku 4.5 (batch processing, classification), Opus 4.6 (newsletter generation).
The numbers: More than 90% of Kindora's codebase was built with Claude Code by Justin and his classmate Karibu Nyaggah. $100,000 raised in year one for his own nonprofit (8 grants, mostly cold outreach using his prototype). $101,000 raised in a single donor month (March 2026). 328 nonprofits onboarded within months of beta. Accepted into 3 accelerators (AWS/Deloitte Social Entrepreneur, Blackbaud Social Good Startup Program, Camelback Ventures); $50K SAFE from Camelback.
Justin's framing of using AI coding tools — "It's like jumping into an Iron Man suit" — has become widely cited. From the Anthropic case study: "There's such a unique opportunity right now for people who understand their problem space."
Read Anthropic's customer story | Listen to the full interview
At enterprise scale (org-wide initiatives)
Salesforce — Zach Stauber
Stauber is the Salesforce employee profiled in HBR's February 2026 article "To Thrive in the AI Era, Companies Need Agent Managers." His formal corporate title is Senior Director, Agentforce Data & AI within Salesforce's Digital Success group. HBR uses the functional descriptor "support agent manager" as a new job category. He sits between corporate strategy and the AI agents doing the work — managing a fleet of generative AI support agents on Agentforce.
The why: Salesforce, like every customer-facing company, has a tier-1 support volume that is largely automatable but also fundamentally consequential. Someone has to own whether the agents are accurate, brand-appropriate, and improving. That's Stauber's job.
What he manages: Configures and refines the agent fleet's prompts and instructions; monitors performance dashboards; reviews the cases agents couldn't resolve; updates the agents as the business changes; defines hand-off protocols (when does the agent route to a human, when does it act alone, when does it escalate).
The stack: Salesforce Agentforce. Prompt engineering as a primary daily activity.
The framing: HBR uses Stauber to argue that "agent manager" is the new product manager of the AI era — translating strategy into agent behavior, managing the agent like an employee.
"Data, Data, Data. I start and end my day in dashboards, scorecards, and agent observability monitoring."
BBVA — Elena Alfaro, Head of Global AI Adoption
BBVA is a major European bank with ~120,000 employees. Alfaro runs the bank's global AI adoption program. She isn't building tools herself — she's building the conditions under which everyone else can.
The why: BBVA's leadership realized that "shadow AI" — employees using ChatGPT and other consumer tools on their own — was simultaneously a security risk and a signal of pent-up demand. Rather than ban it, they built a sanctioned platform.
What was enabled: 20,000+ custom GPTs created by employees on BBVA's ChatGPT Enterprise rollout. Bankers, lawyers, credit analysts building their own assistants.
- A frontline-built Peru ops assistant used by 3,000+ employees cut query handling time 87% (from 7.5 minutes to 1 minute).
- A legal GPT handles 40,000 annual queries with a 9-person team of in-house lawyers.
The stack: ChatGPT Enterprise as the platform. BBVA's own knowledge bases, customer data, and compliance frameworks integrated.
The numbers: 80%+ daily active usage among licensed users. 4,000 GPTs used daily.
Read the OpenAI BBVA case study | BBVA's own write-up
Walmart — nano agents across the org
Walmart, the largest US private-sector employer, has put AI building into the workflow of every team. CEO Doug McMillon's frame: "AI is literally going to change every job."
What's happening: Walmart's agentic strategy uses a layered architecture — "super agents" that preside over specialized sub-agents, including small, rapidly developed "nano agents." Use cases span merchandising, supply chain, customer service, and operations. The premise: the team that owns the workflow is the team that should build the automation.
The framing: Walmart's bet is that the velocity of AI building, not the centralization of it, defines competitive advantage. Many small wins, shipped fast, by the people closest to the work.
Walmart's agentic strategy | CNBC on the McMillon quote
Rakuten — Claude Code adoption across departments
Rakuten, the Japanese e-commerce giant, deployed Claude Code across product, sales, marketing, and finance teams — explicitly extending it beyond the engineering organization.
The why: Non-engineering departments at Rakuten were dependent on the engineering organization for any custom tooling, with quarterly release cycles. Claude Code (and Claude Managed Agents in Slack and Teams) collapsed that bottleneck.
What was enabled: Non-engineering employees across four departments building their own deliverables — spreadsheets, slides, internal apps, analysis tools — without writing code directly. ML Engineer Kenta Naruse is also planning an ambient-agent project that breaks complex tasks into 24 simultaneous Claude Code sessions for monorepo updates, with him providing only occasional guidance.
The stack: Claude Code (terminal). Claude Managed Agents (Slack/Teams integration).
The numbers: 79% reduction in time to market (24 days → 5 days). 97% reduction in critical errors. 99.9% accuracy on complex code modifications. Bi-weekly release cycles, replacing quarterly.
"You can have five tasks running in parallel by delegating four to Claude Code while focusing on the remaining one." — Yusuke Kaji, GM of AI for Business, Rakuten
Read Anthropic's Rakuten story
Stanford GSB — AI@GSB
The largest student organization on the Stanford Graduate School of Business campus by autumn 2025 (597 members). Co-founded by four MBA students: Jenni Steiger (former chief of staff at BlackRock, on Larry Fink's team), Celeste Bean (a hardware and machine learning engineer with 50+ patents from her PlayStation R&D leadership), Abby Alder, and Alfredo Méndez.
The why: Stanford GSB sits next door to the major AI labs but the MBA curriculum hadn't internalized that proximity. The four founders wanted MBAs to use AI daily — not just learn about it. The dean and faculty got out of the way.
What's emerged: ~36 MBA courses that explicitly integrate AI/ML. The Startup Garage now opens with a 60-minute AI-powered hackathon. Students are wiring their own personal-and-team operating systems out of Gmail, Claude, Slack, Notion, and Monday. Steiger and Bean have founded a company together (Steiger Bean, AI strategy for executive leaders).
The hiring market has noticed: summer-internship interviewers now ask MBAs for "portfolios of builds." A student in Jennifer Aaker's class reportedly built an agent to negotiate trucking rates for independent operators and found their way onto the founding team of a logistics-AI startup.
The stack: Claude Code, Cursor, Lovable, the everyday SaaS layer (Slack, Notion, Gmail, Monday). The point is the integration, not any single tool.
The framing (Alfredo Méndez):
"I think everyone is technical now. But not everyone is curious. We are all builders now."
Read the Times of India profile | Poets & Quants on Stanford GSB and AI
What's the pattern?
Look across all twenty-plus examples and three things repeat:
1. The builder isn't a software engineer. A Chartered Accountant, an Operations Manager, a marketer, a clinical pharmacist, an MBA student, a former program officer, a media buyer, a lawyer. Most have never built production software before.
2. The first build is small. A journal posting tool. A late-invoice reminder workflow. A brand-mention tracker. A landing page. A criminal-record check. The wins compound; the first build is rarely the most ambitious.
3. The tool changes; the pattern doesn't. Power Platform, Make, Zapier, n8n, Lindy, Lovable, Cursor, Claude Code, Copilot Studio, Bubble. Different stacks for different jobs. The skill that ties them together is briefing — explaining clearly what you want, iterating until it works, owning the running system.
If this is you, you're an Agenteer. We're building the community for the people figuring this out together at agenteers.ai.
Want to share an Agenteer story? Drop us a note. The list keeps growing.