Get your team shipping AI-assisted code you can actually review.
A remote, full-day workshop for product, engineering, and reviewers. You leave with a governance charter, review gates, and an ownership model your team follows the next morning.
Sessions run remote and are scheduled in your timezone — Europe, North America, and APAC hours are all available.
“A ticket our own engineers had sized at ten days went out the door in about five hours — during the workshop, on our first time using it.”
In the room: CEO, CTO, product manager, lead developer, and three junior engineers. First full-day session, working on a real backlog item from their own queue.
Your engineers adopted AI coding tools without waiting for a rollout plan. The work is already being merged — the governance layer to see it, review it, and answer for it is what's missing.
Agents are writing code across four different tools. Which review actually caught anything?
What shipped this sprint, and how much of it did a human genuinely read?
You spent the budget on AI tooling. Show us the return, with a method behind it.
Name the person who reviewed and accepted this agent-generated change.
of enterprise generative-AI pilots show no measurable P&L impact, according to MIT's NANDA initiative. The tools are not the bottleneck. The operating model around them is — which is exactly what a rollout day is for.
Artifacts written in the session, in use the next morning.
Not a slide deck and a feeling of alignment. Documents your team wrote together, in your own repository, that govern how AI-assisted work gets specified, reviewed, and accepted.
A governance charter
A Charter draft matched to your delivery standards — what agents may do unsupervised, what needs a human decision, and where the boundaries sit.
Review gates for specs and work packages
A concrete cadence for reviewing intent before implementation and accepting work against defined requirements after — so review is a gate, not a rubber stamp.
Stakeholder visibility and privacy posture
What leadership and compliance can see without a status meeting, what leaves your environment, and what does not.
A cleaner ownership boundary
Product owns the what. Engineering owns the how. Written down, with the grey areas named — because the grey areas are where AI-assisted work goes wrong.
A rollout plan across your existing tools
Claude Code, Cursor, Copilot, Gemini CLI — whatever your team already runs. Nobody is asked to switch tools or migrate a tracker.
One vocabulary for plans, work packages, and done
The number-one problem we see in these sessions is not the tooling. It is that four people mean four different things by the same ticket.
Three steps. You can stop after the first one.
A 30-minute fit call
We look at how your team is using AI coding tools today and where review and ownership are breaking down. You get a written rollout sketch back.
The full-day remote workshop
Product, engineering, and reviewers in one room. We write the charter and the review gates against a real item from your backlog, not a toy example.
Quarterly check-ins
We come back each quarter to see what held, what drifted, and what the charter needs to say now that the team has been living with it.
You keep the rollout plan whether or not you book the workshop.
The fit call produces a written rollout sketch for your team — where your review gaps are, what to fix first, and what an operating model for your stack would look like. It's yours to keep and act on internally. If a full day isn't the right next step for you, we will say so on the call.
Robert Douglass
Maintainer of Spec Kitty · facilitates the workshop himself
Robert has spent more than twenty years building and shipping developer platforms across CMS, e-commerce, and DevOps, and maintains Spec Kitty's open-core CLI in the open on GitHub. He runs these sessions personally — there is no handoff to a training partner, and the person who wrote the tool is the person answering your team's hard questions about it.
One price, one day, nothing else to work out.
Full-day remote rollout workshop
One full day, remote, up to 20 participants. Includes one month of the Spec Kitty SaaS offering for the team and quarterly check-ins as follow-up.
One month of the Spec Kitty SaaS offering, so the team can use the shared workflow and enablement features the morning after the session.
Quarterly check-ins to catch drift in the charter and the review cadence once the team has been living with them.
The open-core CLI stays available separately at no cost, under MIT on GitHub. The workshop is about the operating model, not a licence.
The rest of the day's agenda
- Spec review in natural language before anything is implemented
- Work-package planning and kanban discipline across parallel agents
- Keeping decisions and rationale from getting lost between sprints
- Working a real backlog item together, end to end, in the session
- Fitting the workflow to Jira or Linear without abandoning either
- Where the tooling stops — what still needs a human, stated plainly
Questions we get on the fit call.
Request the full-day workshop
Skip the fit call if you already know you want the day. Tell us who should be involved, how large the team is, and the delivery problem you want the session to solve.
Start with the 30 minutes, not the full day.
We look at how your team is shipping AI-assisted code today, and you get a written rollout sketch back — yours to keep whether or not a workshop follows.
Sessions run remote and are scheduled in your timezone — Europe, North America, and APAC hours are all available.
