Everyone has the tools. Claude Code, Cursor, ChatGPT, Copilot — AI coding assistants are available to anyone. But designers and business operations teams don't know what to do with them. Not because the tools are hard to use, but because the concepts underneath — APIs, version control, deployment, system architecture — are foreign. The tools assume knowledge their users don't have.
The gap isn't talent or intelligence. It's literacy. The people who understand the problem domain (designers, strategists, operators) can't build the solution. The people who can build (engineers) don't always understand the problem. AI coding tools can close that gap — but only if someone teaches the concepts that make the tools useful.
People kept asking. Colleagues, clients, external partners — all some version of the same question: "Can you show me how to actually build things with these AI tools?" Not prompt engineering. Not ChatGPT tips. How to build working software. How to go from an idea to a deployed application that does something real.
Bill had been doing this himself for years — designing AI systems and then building them with AI coding tools. The workshop is that process, externalized into a curriculum that non-engineers can follow. It started with internal design teams and expanded to external organizations as independent consulting.
Designers. Strategists. Business operations. Creative directors. Product managers. People who have never opened a terminal, never called an API, never pushed code to a repository. People who have ideas about what AI should do but no way to build it themselves.
These are not engineering bootcamps for career-switchers. These are workshops for people who will keep their current roles but gain the ability to build prototypes, automate workflows, and speak the language of the systems they're designing.
Six modules. Each one introduces concepts that used to require a CS degree. AI coding tools — Claude Code, Cursor — are the bridge that makes them accessible to non-engineers.
The centerpiece of every workshop is a working system that participants can open, inspect, and modify. The current demo is a deck-to-video production pipeline — an 8-step automated workflow that takes a presentation deck and produces a rough video cut through chained API calls.
Participants don't just watch. They open the system, read the code, modify the prompts, swap APIs, add steps, break it and fix it. The demo is a teaching instrument — every intermediate artifact is visible, every connection is traceable. Participants learn by doing, not by watching someone else do it.
"Wow, I didn't realize I could do this."
— The most common reactionThere is a specific moment in every workshop when participants build their first working application and something shifts. They realize it is simultaneously far simpler than they expected — you can build a functional app in a single session — and far more complex than they first thought. The system design underneath is real engineering thinking, real architecture, real decisions about how pieces connect. The AI tool handles the syntax. The human handles the intent.
That double realization — "I can do this" and "this is deeper than I thought" — is the unlock. It's what turns a one-day workshop into a permanent capability. Participants leave not just with a working application, but with a mental model for how AI systems work that changes how they think about every tool they use afterward.
Portrait Media — Workshop co-led with Mickey McManus for Bonin Bough's media organization. Focused on AI production workflows for media teams.
BCG Design Studios — Internal design teams across multiple offices. The original audience — colleagues who watched Bill build AI systems and wanted to learn how.
Enterprise clients — Workshops delivered to design and operations teams at a major global automotive manufacturer, a CPG company, and a federal government client. Curriculum adapted to each organization's specific workflows and tools.
The workshop curriculum became a body of published writing. These articles are the frameworks externalized — the concepts Bill teaches in person, written down so they can travel further than any single workshop.
"From Tools to Systems: Foundations of Agentic System Design" — The Warhol Factory metaphor. How to see your workflow as a system and automate the mechanical parts. The 8-step pipeline. Intent, context, workstations, orchestration, inspection, loops. This is the workshop curriculum in article form.
Additional frameworks on context engineering, agentic system design, and AI production workflows published on Medium and Substack — each one originating from patterns discovered while teaching.
The gap is literacy, not ability. Designers and business people are fully capable of building AI systems. They lack the vocabulary (API, SDK, repository, deployment) and the mental models (systems thinking, chain architecture, spec-driven development). Once those are in place, the AI coding tools do the rest. The workshop doesn't teach people to code. It teaches people to think in systems — the tools handle the code.
The spec is the most powerful prompt. When participants learn to write a clear specification — what the system does, what it doesn't do, what tools it uses, where the human checkpoints are — they discover that the AI coding tool can build most of it. The bottleneck was never the building. It was the describing.
Working software teaches faster than slides. Participants learn more from 30 minutes inside a working pipeline than from an hour of explanation about how pipelines work. The live demo isn't an illustration of the concepts. It is the concepts. Opening the factory and looking inside is the curriculum.
Teaching is R&D. Every workshop produces new insights about how non-engineers think about systems, where the conceptual gaps are, and which metaphors actually land. The Factory metaphor came from teaching. The four-loop model came from teaching. The published frameworks are teaching artifacts — patterns that emerged from watching people learn.