
After the Pilot · for AI companies whose customers are councils, public agencies and non-profit organisations
Your renewal is a conversation your customers are already having. Give them a room to have it.
Four to eight of your customers. Three sessions over eight weeks. Each team writes down, in its own words, how it uses your tool, and you take those words into every renewal.
Four design-partner places, winter 2026 to March 2027.
The stall
The tool still works. The habit didn’t travel.
In Nava Labs’ study of caseworkers at six organisations in Los Angeles County, use fell over time, except where three things were in place:
None of the three ships in a software update. After the Pilot puts them in place, team by team.
What public-sector rollouts show
Five public-sector rollouts. What each did to keep people using it.
17 → 450staff, from pilot to rollout
Managers had their own training on building the tool into their team’s week.
A manager track3 → 10+social workers, then wider
Practitioner feedback, read every week, decided which templates changed and where training went.
One weekly view~7,500users across the state
Champions picked from the earliest users helped colleagues on site and passed feedback back.
Named champions6organisations, 61 caseworkers
Use held where training, organisational backing and peer support were in place.
Peers to turn to29 → 1,100staff, from pilot to rollout
An independent evaluation came before the decision to roll out.
Evidence for the decisionResults belong to the organisations named. Sources: St Helens Council; Local Government Association case study (Kingston); GovTech (Illinois DCFS); Nava Labs (Los Angeles County); techUK and Unity Insights evaluation (Kent).
The eight weeks
Six stops. Who does what at each, and what’s on the table afterwards.
Their teams
You
On the table afterwards
Every cohort is shaped with you. Design partners get the full session plan.
What teams leave with
Three pages that outlast the cohort.
How we use it
- Use it for
- Notes from home visits and reviews
- Not for
- Safeguarding decisions
- Reviewed by
- The practitioner, before saving; the team manager spot-checks weekly
- When it’s wrong
- Correct it, note it in the team log
- Ask
- Our two champions
Agreed at the team meeting · signed by the team manager
A reflection with peers
What I tried. What a colleague at another council showed me. What I’ll do differently with the people I support.
Written in the final session, discussed with peers, kept for their professional development record.
Usage by team
Plus product issues by site. Each customer approves its summary before you use it.
Examples of the format · not a real client or real data.
What your customers are already being asked to show
The team pages, review practice and reflections are records your customers can point to. They support these requirements. They are not a compliance service.
Where it fits
Not another course. The layer between your training and your renewal.
| Free AI courses | Your own onboarding and training | After the Pilot | |
|---|---|---|---|
| Who learns | Individuals | Users inside one customer | Teams, across several customers |
| Tied to | AI in general | Your product’s features | How each team works with your product |
| Who they learn from | Course instructors | Your team | Organisations a few months ahead |
| What’s left behind | A certificate | Trained users | A written agreement per team, and evidence for renewal |
It builds on the champions and managers you’ve already trained. When Nesta reviewed one social care AI tool, it recommended the developer convene councils to share lessons about implementation. After the Pilot does that work, for you.
How we agree on it
We agree on what it’s for before we agree on a fee.
You’ve already paid to win the account. The open question is whether use spreads before the renewal, which is often decided by people who weren’t in the pilot.
What you want
Use beyond the pilot team, at customers whose renewal is coming.
How we’ll know
Two measures you choose. For example: active users by team at day 90, and customers asking to expand.
What it’s worth to you
Set against the value of one renewal, which you know and I don’t.
Who runs it
Hello, I’m Billy Bicket.
I design the part of a rollout that happens between people: who decides how the tool is used, who checks what it produces, who a practitioner asks when it’s wrong. I design and run every cohort myself.
“The part that’s special is the way Billy can open a process to a community of people, listen openly and honestly to their ideas… and then build a project in which people can see themselves.”Marnie Webb · CEO, TechSoup
Why peers
The best guide is often someone a few months ahead.
In health care, teams from different organisations learning together improved on at least one main measure in most of 64 controlled studies. Whether that holds for AI in public services is part of what the design partners and I will test. Here is where I’ve seen peer learning work:
4 of 4
teams, chosen from 181 applications, finished their archives in the Cultural Memory Lab at TechSoup, with the Filecoin Foundation for the Decentralized Web and Gray Area.
7,581
meetups, run by 1,353 local volunteers. Net Tuesdays at TechSoup.
“Billy’s leadership and thoughtful design of the TechSoup–FFDW Cultural Memory Lab went well beyond our public events… Anyone bringing emerging technology to new audiences… should talk to him.”Kaitlin Donovan · Senior Program Manager, Social Impact, Filecoin Foundation and Filecoin Foundation for the Decentralized Web
From recommendations. Results belong to the organisations where they happened. One comparison isn’t proof. Programmes I helped design were backed by Google, Microsoft, USAID, the U.S. State Department, SAP, Cisco, Yahoo, UC Berkeley’s Human Rights Center and the Filecoin Foundation for the Decentralized Web.
Design partners · 2027
It’s early, and I’ll say so.
“He didn’t defend the old plan. He said early what wasn’t working, and he asked the team what we were seeing before deciding what to change.”Elijah van der Giessen · Program Manager, TechSoup, on the Cultural Memory Lab
It may suit you if
- Your customers are councils, public agencies or non-profit organisations.
- At least four have finished a pilot.
- Renewals in the next twelve months matter to your plan.
What we’ll find out together
- Whether customers learn faster from each other than from your team alone.
- Whether managers can give about 6 hours, and champions about 10.
- Whether use at day 90 holds at the level of week eight.
Contact
Tell me where your customers slow down.
Thirty minutes. Bring the customers, the pilots and the renewal dates. If I can help, you’ll have a written proposal with options within two working days. If not, I’ll point you to someone who can.
Four design-partner places, winter 2026 to March 2027.