AI-enabled delivery leadership

Your developers got faster. Your delivery didn't.

Speed up one step and the wait moves to the next. Coding first, then review, then whatever comes after — the queue relocates rather than clears. Running teams that use AI well means finding where the work waits now, and being honest that it still waits somewhere.

Fifteen questions, about ten minutes, a scored report at the end. No cost.

What actually goes wrong

The technology is rarely the hard part

  • The expectation arrives before the capability

    "AI makes you faster" becomes "deliver more", and the pressure lands on a team whose way of working has not changed.

  • More is produced than anyone can absorb

    Output rises everywhere at once. Whatever the team does with it next — check it, integrate it, decide on it — is where the work starts piling up, and automating that step just moves the pile again.

  • Teams are still people

    Cohesion, trust and the path for junior engineers matter as much as they ever did, and are easy to lose while chasing throughput.

  • Your measures stopped meaning anything

    Velocity and story points were a weak signal before. With AI in the loop they are close to useless, and executives still want a number.

Who this is for

You are accountable for more than one team

Heads of delivery, engineering managers and directors, delivery and programme managers, and agile coaches moving into leadership. Anywhere delivery is a system of teams rather than one squad, AI tools are already in use, and somebody above you wants to see what they bought.

Sector does not matter much. The shape of the problem is the same in a bank, an agency and a government department.

The course

Seven modules, each ending in something you can use on Monday

Short videos, then a worksheet or template, then something to do with your team. The durable parts — flow, team design, economics, evidence — are kept separate from the tool specifics, because the tools will change and the thinking will not.

  1. 00

    Start here

    Score your teams and pick one thing to change first.

  2. 01

    See the system

    Map how work actually flows, and where it waits rather than where it is worked on.

  3. 02

    Pace

    Turn faster coding into faster delivery, using WIP, batch size and review policy.

  4. 03

    Trust

    Definition of Done, testing and guardrails for work an AI helped produce.

  5. 04

    Impact

    Choose which AI opportunities are worth running, and test them small.

  6. 05

    Team

    Roles, skills and team design, including the junior pipeline problem.

  7. 06

    Prove it

    Measures that survive scrutiny, and how to report them upward honestly.

What it won't be

Deliberately not a coaching programme

No one-to-one coaching, no reviewing your plans or code individually, and no recurring live sessions. Those are the things that do not scale, and promising them is how this would quietly become a consultancy with a login page.

Member questions get answered, but in recorded lessons everyone can use. There is no community forum at launch either: moderating one properly is a job, and doing it badly is worse than not doing it.

Where this is up to

Being built in the open, and not finished

The course is in development. There is no price, no launch date and no sign-up yet, and I would rather say that than collect your email against a promise I have not worked out how to keep.

What does exist today is free and real: the Health Check, the articles and the videos. The Health Check is also how the course gets shaped — the answers show which problems are common and which are rare, which decides what gets built first.

Free, no account needed. You can opt in to updates at the end — that is the only list there is.