Enterprise AI Corporate Trainer

Everyone has heard of AI harnesses. Far fewer have successfully built one.
This course closes that gap. In two weeks, you build a working harness on Claude Cowork that produces a weekly project status report on its own, pulling live data and formatting it the way the team wants, every time you run it.
You build it from the four pieces every harness is made of: a skill that encodes how the report should be written, context the harness reads from, a tool that connects to Google Drive for live project data, and a deterministic workflow that ties them together so the report comes out the same way every time.
You build on a shared practice scenario, so the mechanics stay clean and no one is debugging messy data while learning. Once the skills are solid, you transfer the approach to a real target of your own.
There is no code. You work entirely inside the Cowork interface, in live sessions where mistakes get caught and fixed in the room. You leave with the working harness, the mental model to read and interpret any harness, and the ability to build the next one yourself.
Stop watching harness demos and build one. In two weeks, you ship a working harness on Claude Cowork you can run and reuse.
Build the whole thing inside the Cowork interface, no code, in guided live sessions.
Run it on demand to produce a formatted weekly project status report.
Define the report's structure, voice, and standard so output isn't generic filler.
See the difference between a one-line instruction and a real, reusable skill.
Give the harness real organizational knowledge: team roster, project descriptions, prior notes.
Watch the skill produce sharper output once it has real material to draw on.
Wire up Google Drive as a tool and pull live project data straight from a Sheet.
See why a tool differs from context: it reaches outside the harness to act.
Tie skill, context, and tool into one fixed sequence that produces the report repeatably.
Test it, spot failure modes, and confirm it holds up on repeat runs.
Use the three signals (volume, hops, reading) to pick a real target after the cohort.
Apply the same four-piece build to a workflow that matters to you.

Enterprise AI Educator/Enablement (NBCUniversal, National Grid)

Product managers who can scope a use case and run a pilot but have never built a harness, and want to ship their first one.
Analysts and ops people who do recurring reporting by hand and want a harness that produces it, with no coding.
Enablement and L&D leads who teach AI tools and want to build a harness themselves before rolling one out to a team.
Cowork requires a paid plan, and you will use it in every session. A free account will not work.
You connect Google Drive so the harness can read a course-provided Sheet. Personal accounts avoid the blocks common on work accounts.
Everything is built in the Cowork interface. No coding required; if you can use a web app and a spreadsheet, you are ready.

Live sessions
Learn directly from John Francis in a real-time, interactive format.
Lifetime access
Go back to course content and recordings whenever you need to.
Community of peers
Stay accountable and share insights with like-minded professionals.
Certificate of completion
Share your new skills with your employer or on LinkedIn.
A working harness you build and keep
You finish the cohort with a real, running harness on Claude Cowork, not notes or slides. It produces a weekly status report on demand, and it's yours to keep, run, and adapt.
Build in Claude Cowork, no coding
Every step happens inside the Cowork interface. You author the skill, configure context, connect the tool, and define the workflow through the UI, with no code at any point. Mixed technical backgrounds welcome.
Live build sessions with help in the room
Four live working sessions of about 90 minutes, where you build alongside the cohort. First-attempt mistakes get caught and fixed on the spot, which is exactly where a solo learner would otherwise stall and quit.
Connect real data through Google Drive Description
You wire up Google Drive as a tool and pull live project data straight from a Sheet, so the harness works from current data instead of stale copies. A real connection, not a simulated one.
A mental model to read any harness
Beyond the one you build, you leave able to read any harness: the three levels (chat, agent loop, harness), the three primitives (skill, context, tool), and the two operations (discoverability and orchestration). No harness stays a black box.
Student demos in the final session
The last session ends with the cohort demoing what they built. Demos push you to finish, and seeing how others approached the same build deepens the learning more than another lecture would.
Maven Guarantee
Your purchase is backed by the Maven Guarantee.
4 live sessions • 8 lessons • 5 projects
Aug
11
Aug
13
Aug
18
Aug
20
Live sessions
3 hrs / week
Two live build sessions each week, about 90 minutes each. Recorded for replay. The final session ends with student demos.
Tue, Aug 11
7:00 PM—8:30 PM (UTC)
Thu, Aug 13
7:00 PM—8:30 PM (UTC)
Tue, Aug 18
7:00 PM—8:30 PM (UTC)
Thu, Aug 20
7:00 PM—8:30 PM (UTC)
Projects
2-3 hrs / week
Each session has a hands-on lab where you build your harness in Claude Cowork one piece at a time: a skill, context sources, a tool, and a workflow. By the final week you have a working starter harness, then you begin building one on a real piece of your own work for the capstone demo.
Async content
1-2 hrs / week
Eight short self-paced modules, two before each live session. They cover the concepts behind a harness and set up each build step, so live time is spent building, not catching up. Most include a lab. Complete each before its session.
Maven for Teams
Reimbursement
Get your company to pay
Everything L&D needs: email template, receipts, and certificate of completion.
Get reimbursedPrivate cohort
Run a cohort for your org
A dedicated cohort with a custom schedule and curriculum, tailored to your team.
Book a private cohort