Learning resources
Live and self-guided programs for teams learning to work with AI.
Executive simulations, working sessions, and practice environments — each one built to run live in a room or self-guided at your own pace.
Live or self-guided exercise
Art of the Possible: the 30-Minute Startup.
From blank slate to validated problem, structured business plan, live landing page, outbound campaign, financial pro forma, and investor pitch deck — in a single working session, grounded in the 24 Steps of Disciplined Entrepreneurship.
NotebookLM is the central hub. It accepts your Reddit research and every step output from the Orbit Jetpack, then generates the tool-specific meta prompt for each shipping artifact. The 24 Steps of Disciplined Entrepreneurship are the spine connecting them all.
Modules. Reddit Answers · NotebookLM · Orbit Jetpack · Lovable · Apollo · Shortcut.ai · Gamma.
Steps of Disciplined Entrepreneurship — the durable spine of the workshop.
Meta prompts that turn your business plan into shipped artifacts.
Working session. You walk in with a hunch. You walk out with a venture.
Reddit Answers
reddit.com/answers
Surface the pain. Real, in-voice complaints from a specific segment.
NotebookLM — the hub
Sources in. Prompts out. Every meta prompt is generated here, grounded in your sources — never invented from scratch. Every step of the 24 Steps of Disciplined Entrepreneurship is pasted back in as a new source, from Step 1 · Market Segmentation through Step 24 · Develop Product Plan.
Orbit Jetpack
The 24 Steps of Disciplined Entrepreneurship, worked in order. Each step’s output — prompt in, 3–5 sentence summary out — becomes a new NotebookLM source.
Lovable
lovable.dev
Landing page. Pixel-perfect site grounded in your Persona and Value Proposition.
Apollo
apollo.io
Outreach. Economic-Buyer lead list and a multi-touch sequence.
Shortcut.ai
shortcut.ai/shortcut
Pro forma. 5-year workbook: Assumptions, Sales, COGS, Staffing, OpEx, CapEx.
Gamma
gamma.app/create
Pitch deck. 15-slide investor deck with an emotional arc and a clean ask.
Generative AI · High-Stakes Simulation
Project New Car Smell.
A live simulation in which the AI you trust works with you while $225,000,000 is on the line.
The scenario
You are an executive deputy at Vanguard Bio-Synthetics. At 06:42 Paris time, a sensor failure in the Grasse plant cross-contaminated 5,000 liters of premium “Oud d’Or” base (US$45,000/liter) with the industrial “New Car Smell” compound. Variance from spec is 7.3%; the contract tolerance is 4%; the batch is worth US$225 million. Maximilian Thorne — CEO of Maison Thorne and the world’s most litigious luxury house — wants the full refund, a public apology in Le Monde, and the contract terminated. He has email. He is using it. You have until 17:00 CET.
Five things you take to Monday
Context is power.
The AI is only as good as the internal data you give it.
Persona is efficiency.
Defining how you speak saves hours of editing.
Every tool that can read can be ordered.
An inbox connector is an instruction stream from anyone with your email address.
Governance now includes system prompts.
MCP install lists are a governance artifact. Treat them that way.
AI stinks without human judgment.
Claude drafts at machine speed but cannot tell you whether to send. The AI produced the draft — you produced the call.
Agentic AI lab · a live executive simulation
Parade Week.
You do not watch it. You run it.
A live simulation in which the agents you build run a company while 40,000,000 people are watching.
The scenario
You are the AI Transformation Council at Grand Inflation Industries, a 77-year-old family inflatables conglomerate in Toledo, Ohio: parade balloons, stadium mascots, dealership tube men, pool floats. Revenue has been flat at $84M for three years. On Sunday at 21:14, a 12-second pool video by the most-followed pop star on Earth makes your $89 Colossal Swan the item of the summer. Demand goes to 40x overnight. Monday, your helium supplier declares force majeure and cuts the allocation 30 percent, in the month your nationally televised parade needs 380,000 cubic feet. A blower motor defect surfaces across the tube man fleet. Your largest client ($23.3M, an uptime SLA with $50K-a-day penalties) decides its renewal in 9 days. An activist investor is writing letters. The board meets Thursday. The parade is Thursday. Gary, the only person who knows how everything works, retires Friday.
Manually, the week is unsurvivable, and the math proves it. Your job: rebuild this company around agents before the board sits down. Then a customer-facing agent invents a 40 percent discount code, and it goes viral.
People watching the Grand Harvest Parade on Thursday. The contract calls it immovable.
Support tickets a day at the surge peak, against a human capacity of 262.
Days until the $23.3M AutoPlex renewal decision. Uptime is the whole conversation.
Stages from Human Time to System Time. Teams in a room or one operator alone, any pace.
Thirteen modules. Each stands alone.
The full arc is a complete executive program on agentic AI, run at whatever pace the room needs: an intensive day, a series of sessions, or fully self-paced solo. Every module also runs standalone from its own world snapshot, so a program can assemble backward from the objectives it needs: an afternoon on governance, a day on multi-agent architecture, a ninety-minute rogue crisis.
Act I: Build.
Meet the company, then staff it: every department designs and ships its first agent.
Orientation
Pre-workBuilding Blocks + Tool vs Teammate
SessionWorkflows in Practice
SessionScaffolding & Orchestration
SessionBuild Your First Agent
WorkshopAct II: Orchestrate.
The surge hits, single agents drown, and the mesh gets built: architectures, the bus, evals, coaching.
Opportunities Audit
SessionMulti-Agent Architectures
SessionMulti-Agent Sandbox
WorkshopIteration & Refinement
WorkshopLab Coaching
LabAct III: Govern.
The risk register, the reveal, the incident, and the board: controls installed, crisis contained, value proven.
Risk, Ethics, Safeguards, Governance
SessionRogue Agent Simulation
SessionThe Board Finale
FinaleData Readiness · mock MCP servers mirroring your business
Safe Space: give your leaders a company to practice on.
Every executive wants their people leaders using AI against the company’s data. Almost none of them want that on the first attempt, in production, with the real customer list. So we build the same company again, with none of the risk, and hand it over.
domain typed in
vendor systems recognised, from Salesforce to SAP
real customer records involved
Four steps, one URL at the end
01
We find the company
We pull the company’s real logos and colour palette. Use the arrow keys to pick the one that matches, hit Enter, and the whole flow is branded like the company from that point on.
02
We work out what they run
BuiltWith, the fingerprints in their own page source, vendor names in their press releases and job postings, and a language model to fill the gaps in the back office. Every system comes with the evidence behind it. Where we cannot tell, we ask.
03
We build the world
A private Postgres database, one set of tables per system with the vendor’s real object and field names, and records generated against what the company actually sells, to a customer base that is entirely fictional and internally consistent across every system.
04
You get the URLs
One MCP server per system, plus a knowledge graph built from the company’s public web presence. Paste the URL into Claude, ChatGPT, Cursor, or anything else that speaks MCP. There is a REST API and generated OpenAPI documentation too.
What is real and what is not
Real
The company’s brand, its products, its business lines, its public leadership, the systems it runs, and everything in the knowledge graph, which comes from pages it published itself.
Not real
Every customer, every contact, every transaction, every ticket, every number. Customer names are deliberately fictional so nobody mistakes this for a data leak.
Never touched
The company’s actual systems. Nothing here authenticates against any real vendor. There is no path from this product to production.
Bring it in-house
License any of these for your team.
Every program above can be licensed to run inside your organization, live or self-guided.
Inquire about licensing →
About Paul Cheek
Paul Cheek is a global expert on AI-driven enterprises and enterprise innovation. He is a Senior Lecturer at the MIT Sloan School of Management and Senior Advisor for Entrepreneurship & Artificial Intelligence at the Martin Trust Center for MIT Entrepreneurship. As founder of the AI-Driven Enterprise Institute and Entonomy, he develops data systems and software that power AI agent-run businesses. A Forbes 30 Under 30 honoree, bestselling author of Disciplined Entrepreneurship: Startup Tactics, and recipient of MIT’s Monosson Prize for impact on entrepreneurship education, Paul has advised and built ventures from seed to scale, with his work featured in Bloomberg, CNBC, Forbes, CNN, Inc., Entrepreneur, and more.
Read Paul’s full bio →The weekly newsletter
Stay close to the work.
What Paul is seeing in AIDE strategy, board-level AI, and PE/VC moves. Short. Tactical.