1. How old is the core system that holds your case record?

Age itself is not the problem. Age plus no supported interface is what turns a six-week integration into a nine-month one.

Under 5 years, actively supported5 to 15 years, still under a maintenance contractOver 15 years, or the vendor no longer supports this versionWe are not certain which version we run

2. Can another system read your case data programmatically today?

Nearly every useful project depends on this answer, and it is the one most often assumed rather than checked.

Yes, a documented interface we have usedYes, but only through the vendor, at a cost per requestOnly by scheduled file exportNo, or nobody here knows

3. Is there a current written inventory of the systems your agency runs?

Without it, every project begins with four weeks of discovery that somebody has to pay for.

Yes, maintained and currentYes, but more than a year oldIt exists in people rather than documentsNo

4. If you pulled 100 case records at random, how many would have the fields you need, correctly filled?

This is the single best predictor of whether an AI project or a migration will disappoint. It is also the question most agencies have not yet measured.

Nearly all of themMost, with known gaps in specific fieldsFewer than we would like, and we know itWe genuinely do not know

5. Who decides what a data field means when two divisions disagree?

Data governance sounds like an abstraction until a migration forces the question, at which point it becomes a schedule risk.

A named person or standing group with authorityIt goes to IT, who make a judgment callIt is resolved case by case, informallyIt has never come up, or it stalls

6. Can you answer a question like "how long does a case of this type take" without a special request to a vendor?

If the answer is no, you cannot measure whether any project worked.

Yes, our own staff can produce itYes, but only one person knows howOnly by asking the vendorNo

7. Does your agency have an adopted position on staff use of generative AI?

Staff are already using these tools. The question is only whether the agency has said anything about it.

Yes, adopted, with an approved-tool listYes, adopted, but no operating detail underneath itA draft exists, or it was borrowed and not adaptedNo position

8. Who owns the AI question at your agency?

Programs with a named owner tend to go well. Programs without one become everyone's second priority.

A named executive whose job includes itA committee that meets on a scheduleIt sits with IT by defaultNobody, yet

9. If a division wanted to start using a new AI tool next week, what would happen?

The answer describes your real governance, whatever the policy says.

A defined vetting process they would have to passAn informal check with IT or legalThey would ask forgiveness laterThey would just do it, and nobody would know

10. Who would maintain a new system after it is delivered?

The most common failure in court technology is a good build that nobody was resourced to maintain.

Our own technical staff, with capacity for itOur staff, but they are already fully committedA vendor, under a support contractIt has not been decided

11. When a system changed recently, how was the staff prepared for it?

Training is usually the first line item cut, and the one agencies most often regret cutting.

Role-based training, before go-live, with time protected for itA session or two, fitted around the workWritten instructions circulatedPeople worked it out

12. How did the last significant technology project here end?

The last project sets what your staff expect from the next one, and it is a better predictor than any plan.

Delivered, adopted, and still in useDelivered late, but usedDelivered, but people work around itAbandoned, or never fully worked

Find out what your agency can support today

What happens to your answers

Your result

By dimension