AI Implementation
We build a working pilot to your adopted AI policy, with a person reviewing every decision that touches a case, and measure it against criteria agreed in advance.
Where this work starts
Guidance on court AI is now widely available. What is usually missing is a team to build the system. Your policy is adopted, a use case is obvious, and the gap between a workshop and a working system is where most court AI programs stall.
Your agency also cannot afford a public failure. Whatever ships has to keep a person in the loop wherever the work touches a case, log what it did, and be explainable to a judge who was not involved in the project.
- Written for
- The AI executive
- The operator
- IT and security
What we do about it
- Scope
- one workflow, with the acceptance criteria written before the build starts
- Build
- the pilot against your policy, with human review at every decision point
- Instrument
- it, so you can see what it did, how often, and where it was wrong
- Run
- it alongside your staff for a defined period, and measure it
- Hand over
- the code, the prompts, the evaluation data, and the runbook
We recommend, draft, build, and operate. Your agency decides, adopts, and approves.
How the engagement runs
Each phase names what we need from your team and ends at a decision that belongs to your agency. Stopping at a gate is a normal outcome, and the work done up to it is still yours.
Scope
Week 1- From your team
- Process owner and one subject expert
- Decision gate
- Written scope and acceptance criteria you approve
Build
Weeks 2 to 5- From your team
- Weekly review, 60 minutes
- Decision gate
- Demonstration against the acceptance criteria
Supervised run
Weeks 6 to 9- From your team
- The staff who will use it
- Decision gate
- Measured results reviewed with you
Handover
Week 10- From your team
- IT takes delivery
- Decision gate
- You decide whether it continues, expands, or stops
What you get
- Working pilot in your environment
- Acceptance criteria, agreed before the build
- Human-in-the-loop design, documented decision by decision
- Evaluation results against those criteria
- Audit log and disclosure text for the public where the tool is public-facing
- Runbook and handover, including the source and the prompts
What your agency owns at the end
The code, the prompts, the evaluation data, and the runbook are yours. Run it yourself, have us run it, or stop it: the pilot is scoped so that stopping is an acceptable, low-cost outcome.
Deliverables are handed over in editable source. There is no license to renew, no exit fee, and nothing that stops working if the engagement ends.
What it costs
This line is bought through the engagements below. Each one publishes a committed starting price, and its page lists what moves the price above that starting point.
Pilot
We build one workflow under your adopted AI policy and run it with your staff. Acceptance criteria are agreed in writing before work begins, and at week 10 your agency decides whether to continue.
- Price
- From $35,000
- Timeline
- 10 weeks
The safe first step
The readiness assessment is free, takes about 20 minutes, and returns a scored report you keep. It creates no obligation, no call, and no follow-up.