AI Services · Equip
AI training on your work, not someone else's slides.
Hands-on AI training on your own code and workflows, taught by people who run agents daily and measured before and after.
Licenses don't make people productive
Generic courses are free or cheap, and they rarely change how a team works on Monday. In BCG’s 2025 AI at Work study, only 36% of employees felt adequately trained. Employees with five or more hours of training, especially in person with coaching, were significantly more likely to become regular users (BCG, 2025).
The difference is practice on real work, with someone who has done it beside you.
What you get
Leadership briefing
half day · fixed fee
What AI can and can’t do for your business today, the risks that matter, and the decisions leadership needs to make.
Team lab
1 day · up to 25 people · fixed fee
A hands-on day on your own repository or workflows, with exercises built around your real tasks.
Pair Sprint
2 weeks · fixed fee
We pair with your engineers on real tickets with our agent harness running on your repository.
Champions program
6 weeks · fixed fee
Your internal champions learn to run the method themselves, with a recorded curriculum they keep.
Operations track
4 sessions · fixed fee
Productive, safe use of desktop agents and assistants for non-engineers: what never to connect, how approvals work and how to check the output.
Taught on real work, by people who do it
Your work, not a textbook
Exercises use your code, your documents and your workflows, or synthetic copies of them, so the skills transfer the next morning.
Practitioners, not presenters
Our trainers run agents in daily delivery under the same controls they teach.
Measured
Pair Sprints report lead time, review latency and cost per merged pull request before and after, so you see what changed.
Training lands best on a live system
Training works best alongside a harness install, a governed platform or a new agent, so people learn on the system they’ll use. Every format also runs on its own, and we sell training by the engagement, never by the seat.
What we teach is what we run
~480
rules for AI coding agents our trainers teach from, in daily use across our engineering estate
81
blind-scored test runs showed rules about one-eighth the length work as well as full versions, so we teach the short ones
18
formal challenges our agents filed instead of quietly changing a test, the habit we teach
Questions team leads ask
Which AI tools do you train on?
The agents and assistants your teams use, including Claude Code, GitHub Copilot, Cursor, ChatGPT and desktop agents.
Is our code or data used in training material?
Only inside your session, and only if you choose. We can build every exercise on synthetic copies of your workflows instead.
Can non-technical teams take part?
Yes. The operations track is built for finance, HR, customer service and operations staff using assistants and desktop agents.
A free way to start
Take the AI Engineering Scorecard: 15 questions and a gap report that shows which skills and habits your teams need first.
Plan a lab on your own work.
Tell us which teams, which tools and what you want them doing differently in a month.