AI fluency for working teams
A four-week program that ends with every person on your team using AI in their actual job — not knowing about it.
Everyone knows they should be using AI. Almost nobody has been shown how.
Not in the context of their own job, on their own work, with someone in the room who can tell them when they've got it wrong.
So they avoid it entirely, or they use it badly — pasting things they shouldn't, trusting answers they haven't checked, quietly getting worse results than they'd get on their own. Meanwhile leadership assumes adoption is happening.
Watching a video doesn't fix this. Neither does a one-hour lunch-and-learn.
A live cohort for one team inside one organization.
Ninety minutes a week for four weeks. Every participant works on their own real tasks the whole way through, and every participant finishes with one working AI-assisted workflow they use in their actual job.
Not a webinar. Not licenses you buy and hope people finish. A live cohort with an instructor who has already done this at scale.
How it works, and where it fits your job
A working mental model, no math — including why the tool sounds confident when it's wrong. Then each person maps their own recurring tasks and sorts what AI should touch and what it shouldn't.
Asking better
The difference between a question and a brief. Practiced on their real work, not sample exercises. Everyone picks the workflow they'll build.
Checking the answer, and doing it safely
How to catch fabrication, and what "good enough to send" means when the stakes change. Then what never goes into a prompt, and how policy turns into daily decisions.
Show and tell
Everyone presents what they built. Live troubleshooting. Leadership is invited to this one.
Each person
- One working AI workflow from their own job
- A list of their tasks sorted by what AI can help with
- A habit for checking output before it goes anywhere
The organization
- Attendance and completion records
- Before-and-after self-assessments
- A consolidated view of what your team is ready to use
- A one-page readiness summary for leadership
That last one matters. Most training ends with a feedback form. This ends with a document your leadership can act on.
Teams of 10 to 25 who work with information and documents.
Finance, operations, HR, claims, underwriting, compliance, marketing, administration — anywhere the work involves reading, writing, deciding, and sending.
Not built for engineering teams wanting technical AI training, and not built for organizations who want one awareness session.
Remote by default.
Ninety minutes a week is an easy thing to put on a calendar. A half day isn't. People work on their own machines with their own tools and their own files — which is the point, since they're building something they'll keep using.
The in-person option
Two of the four sessions work better in a room: the kickoff and the final show-and-tell. The kickoff turns fifteen individuals into a cohort. The final session is where your people demonstrate what they built, and it's worth having leadership physically present for it. The middle sessions stay remote either way.
Pennsylvania employers: this is likely reimbursable through WEDnetPA, the state's training grant program. Eligible companies can recover up to $2,000 per employee. At either price, a fifteen-person cohort sits inside that ceiling. Your regional WEDnetPA partner confirms eligibility — happy to help you find them.
Brian Harvey spent his career building training for people who didn't ask for it and didn't have time for it.
Most recently as Global Head of Learning & Development for Technology and Engineering at Chubb, where he built and ran the AI training program — 4,000 engineers trained, inside an organization of 7,000. Before that, a global software engineering academy reaching 2,800 engineers across four regions.
That was a heavily regulated, risk-averse, data-sensitive environment. The kind where legal asks hard questions and nobody gets to move fast and break things. He knows where these rollouts stall because he's already hit every one of those walls.
The difference isn't the curriculum.
Basic AI content is everywhere and getting cheaper. What isn't available off the shelf is someone who has already run the rollout you're about to attempt.
Certified by Anthropic in AI Fluency and Claude Code in Action · MBA, Penn State Smeal · UNC Kenan-Flagler Talent Management Institute
Also available: talent strategy, succession planning, mentorship program design, and apprenticeship and early-career pipeline work. Engagements are supported by a network of senior practitioners where scope calls for it.
A twenty-minute call is enough to know whether this fits.
No proposal required to start. If it isn't a fit, I'll tell you on the call.
Email hello@grainfield.ai Or reply to whoever sent you this page — they can make the introduction.