Teach your team to build with AI. Not just use it.
A hands-on curriculum that moves your people from asking ChatGPT questions to building workflows and agents on your own processes. Six working sessions. Every one ends with something they built, not notes.
McKinsey, The State of AI (2025). Rating from post-workshop feedback, September 2026.
Your people are consuming AI. The 6% are creating with it.
Most companies bought licenses and got a room full of people using ChatGPT like Google. That is not an AI plan. That is a subscription.
- Uses ChatGPT like a search engine
- Gets a different answer every time, so nobody trusts it for real work
- Pastes customer data into free tools
- Ten people doing ten slightly different things, none of it shared
- “We use AI” means “we have a subscription”
- Moved one real process from a chat window into a saved, repeatable workflow
- Wrote down how the business works so AI reads it before every task
- Knows exactly what it is good at and where it lies
- Built the first agent that does work while they do something else
- “We use AI” means “the Monday report builds itself”
The difference is not the tool. Everyone has the same tool. The difference is whether anybody taught your people to build with it.
Every session moves a person one rung up this ladder
On a process they actually own. Not a demo. Not an example. Theirs.
Assistant
You chat with it. Everyone is already here, and no money lives here.
Workflows
It runs a repeatable process the same way every time, whether or not the owner is in the room. This is where the 6% separate from the 88%.
Agents on your data
It monitors, decides, and acts on one job while your people do something else. Built by them, checked by them, owned by the company.
“I would pay for sessions like these, especially building on the prompts we learned. It is great to take out time to do these things, rather than trying on your own and finding time during the day.”
Every attendee who left feedback named the specific thing they were going to build next: a scheduling prompt, a process out of their head and into a written SOP, a customer update that sends itself. That is the whole point.
Six working sessions. Nobody sits and listens.
Delivered on your site, on your laptops, on your processes. Every session ends with an artifact in someone's hands.
Stop Learning About AI. Start Using It.
AI 101 that respects your time, then four exercises on your real work: the blind spots audit, the second opinion, the 10-minute SOP, and write it once.
A written SOP, a saved reusable prompt, and a 30-day plan
The Customer Never Waits Again
The gap between a customer reaching out and a human answering is where every company loses money. Close it with response systems written in your voice.
A saved response system for the inquiry that costs you the most
Get It Out of Your Head
Every process that lives in one person's head, documented in 30 minutes. Then the company context file every future AI conversation reads first.
Documented processes and a company context file
Sound Like You, Get Found by Them
A voice guide built from your own best writing. One asset turned into a month of content. Then the AI visibility audit: what ChatGPT says when someone asks who to hire in your category.
A voice guide and a readout of what AI currently says about your company
Your Numbers, Your Answers
Your own data, live. The Monday report that builds itself. What not to automate, and how to evaluate an AI vendor without getting sold.
A one-page AI plan: one workflow, one owner, one number it moves
The Executive Agent Lab
Four hours, leaders and operators in pairs. Write the delegation brief, build version one on real examples, connect one approved source, break it on purpose, fix it, demo it. Then run it for 30 days.
A working prototype, its delegation brief, test results, and a 30-day pilot with a named owner
Why the order matters
Session 6 is the Executive Agent Lab, and it is also available on its own for a leadership team. The difference is what you walk in with. A cohort arrives with the SOPs from Session 3 and the context file from Session 5 already built. That is exactly the prep a standalone Lab has to do from scratch. Same method. The cohort just gets further.
How it is built
Five rules. They are what keep a room of 20 people building for two hours.
- Nobody leaves with notes.
Every session produces an artifact: an SOP, a saved prompt, a context file, an agent. If people only wrote things down, the session failed.
- Role tracks in every exercise.
Operations, sales and marketing, admin and front office, leadership. Everyone works on their own job at the same time.
- Your processes, not examples.
Never “a business.” Your quote-to-invoice handoff, your month-end close, your new hire's first week.
- Guardrails built in.
What it gets wrong, what never goes into a free tool, what needs a human signature. Verify anything touching money, safety, or the law.
- Compounding, not standalone.
Same cohort, six sessions. Session 3's context file is what makes Session 6's agent work.
Three ways to run it
Every format includes the deck, printed prompt cards, a mobile tap-to-copy prompts page, and a feedback loop after every session.
The Kickoff
One two-hour hands-on session for up to 40 people. Session 1 of the curriculum. The way most organizations start.
The Program
Six sessions over six months for a cohort of 12 to 20. Monthly cadence so people ship something between sessions.
Train the Trainer
The program plus the facilitator guide, prompt packs, and deck, so an internal owner runs it for the next cohort without us.
Who is teaching this
Matt Sitek. Twenty years of Fortune 500 technology consulting, now building AI agents for operating businesses in Metro Detroit. Two hands-on workshops delivered for the Home Builders Association in September 2026, and a five-session series for a Macomb County chamber of commerce starting in November. The Level 3 session is honest instead of a sales pitch because Rivet builds Level 3 for a living: a quote engine that took one contractor from 60 minutes per estimate to two, and a phone agent that booked $60,000 of storm work over one weekend that would have gone to voicemail.
Questions people ask before they book
What is a creator cohort?+
A creator cohort is a group of 12 to 20 people from one company who go through six hands-on AI sessions together. Each session is a working session on their real processes, not a lecture, and each one ends with something they built: an SOP, a saved prompt, a context file, a response system, or an agent. The goal is to move people from using AI like a search engine (consuming) to building repeatable workflows and agents on their own work (creating).
Who is this for?+
Organizations with 50 or more people, usually in Metro Detroit, where leadership has been asked what the AI plan is and the honest answer is a software subscription. The buyer is typically a COO, VP of Operations, HR or L&D lead, or an owner. It is not a course for individual consumers.
Do participants need to know how to code?+
No. Every session runs on the AI tools your company already has (ChatGPT, Claude, Gemini, or Copilot). Participants write instructions in plain English, build on their own examples, and test the result. Nobody writes code.
How is this different from an AI keynote or a lunch and learn?+
Nobody sits and listens. Laptops are open inside the first 15 minutes, every exercise runs on the participant's own work, and every session produces a finished artifact. A keynote leaves people inspired. A cohort leaves them with a written SOP and a running workflow.
What does the Executive Agent Lab produce?+
Four working agent prototypes, each built by a leader and operator pair on a real recurring process, plus a delegation brief, recorded test results, and a 30-day pilot plan with a named owner, start date, review date, and stop condition. The Lab is the capstone of the six-session program and is also available on its own for a leadership team.
How long is the program and where is it delivered?+
Six sessions, typically one a month, delivered on site at your location in Metro Detroit. Each session runs two hours except the Executive Agent Lab, which runs four. Organizations often start with a single two-hour kickoff before committing to the full program.
Find out what your team could build in six sessions.
Thirty minutes. You describe the team and the processes that eat the most time. You leave knowing whether a kickoff or the full program is the right first move.
Book a scoping call