Curriculum
On the first morning we put ChatGPT, Claude and Microsoft Copilot through the kind of paperwork your office deals with, and you sort your own information into what can go into them and what can't. The second morning is for building three workflows on your own work, then writing the policy and the plan that keep them running.
Day 1: what the tools can do, and what your data allows
3 hours, live
We start with what the tools are good at
ChatGPT, Claude and Microsoft Copilot go up on the screen, and we give them the kind of work an office does all week: drafting a reply, summarizing a long email thread, pulling the numbers out of a document. You'll see where each one holds up and where it makes things up with a straight face, and we'll show you the checks we run before trusting any of it.
Then we run them on paperwork that looks like yours
This is the kind of material we bring: a contractor's bid, a property manager's tenant emails, a job-costing spreadsheet that's been passed around too many times, an engagement letter from an accounting firm. You watch what comes back, mistakes included, and you'll probably spot some of the problems before we point them out.
Last, you build your data map
This part takes the most thought. You go through what your business handles, from customer records and contracts to payroll and anything covered by a client agreement or a regulation, and put each one in a column: fine for cloud tools, keep in-house, or not decided yet. We'll also show you where to look in a tool's terms to see what it does with your data.
You leave with: Your data map, which decides what you build with on day two.
Day 2: building on your own work
3 hours, live
You build three workflows on your own tasks
Most of the morning goes here. You pick three things your company does every week and build an AI workflow for each, using only what your data map says is fine. Typical examples are turning site notes into a client update, drafting the weekly crew schedule, or replying to the warranty question that comes in every few days, and each one gets tested on a job of yours before we finish.
You leave with: Three workflows, each tested on your own work.
Then you write the policy your staff will follow
It's one page that says which tools are approved, what never goes into them, when a person checks the output, and who to ask. Because it comes from your data map, the rules fit your company and not somebody else's. Your staff are probably using some of these tools already, so it helps to hand them the page sooner rather than later.
You leave with: A one-page AI use policy you can hand out on Monday.
And you plan the next 30 days
Each workflow gets an owner, a start date and a check-in a few weeks out. We keep the plan small enough that a company with a full workload can stick to it, because the point is that you're still using these workflows a month from now.
You leave with: A 30-day plan with a name next to every workflow.
In person
The in-person workshop covers both days over a full weekend in a host city. Here's what changes.
- Rack
- There's a private AI rack in the room, and you'll load your own documents onto it, including the ones your data map kept away from cloud tools, to see how a model running locally handles them and where it falls short. Documents attendees load onto the on-site rack are deleted at the end of the weekend, before the hardware leaves the venue.
- Time
- You get a full weekend instead of two mornings, which means more time on your own workflows with one of us sitting next to you.
- Second seat
- Bring whoever's going to own the rollout. It's usually easier to run a plan you helped write.
- Recording
- In-person workshops are not recorded. In-person registrants receive the workshop materials and 12-month access to the most recent virtual cohort replay.
- Dates
- None scheduled yet. Organizations can ask to host one.
See it before you sign up
The free live session is a preview of the first morning: the tools on real paperwork, and the start of your data map.
The first virtual cohort runs in the first quarter of 2027, and we'll email the waitlist when dates and prices go up.