The credit union AI readiness self-assessment

Twelve questions across five pillars. Answer honestly; a flattering score helps nobody. Your results include a band, a pillar breakdown, and a suggested next step.

Scoring model

Each answer scores 0 to 3 points. Maximum score: 36.

Band	Score	Meaning
Early	0 to 12	AI activity is ad hoc or absent. Foundation work needed before any deployment.
Developing	13 to 22	Pockets of progress with gaps an examiner would find. Governance and data work should come before scaling.
Established	23 to 30	Solid foundation. Ready for structured pilots in back-office workflows.
Advanced	31 to 36	Ready to scale, formalize an internal AI enablement team, and evaluate member-facing uses.

The questions

1Does your credit union have a written AI strategy or roadmap approved by senior leadership? No, and no plans to create one It has been discussed but nothing is written A draft exists but is not approved or funded Yes, approved with budget and owners

2Who owns AI decisions today? Nobody; vendors approach whoever answers IT handles it informally A named executive sponsor without dedicated time A named owner with time allocated and board visibility

3Has the board been briefed on AI risk and opportunity in the last 12 months? No Once, informally Yes, with a written summary Yes, on a recurring schedule with a standing agenda item

4How would you rate the quality and accessibility of your member and transaction data? Scattered across systems, hard to extract Extractable with significant manual effort Centralized warehouse or lake covering most systems Centralized, documented, and refreshed on a schedule

5Do you have documented data classification (what is sensitive, what can leave which systems)? No Informal understanding only Documented but stale or partially enforced Documented, current, and enforced with technical controls

6Do you have a written AI or generative AI use policy for employees? No A verbal "be careful" guideline A written policy, not yet trained or enforced A written, trained, and enforced policy with a review date

7Does your vendor due diligence process include AI-specific questions (training data, model updates, data retention, fourth parties)? No We rely on the vendor's standard security packet Some AI questions added informally A standing AI addendum in every relevant vendor review

8Could you produce a list of every AI-powered system in use at your credit union today, including features inside existing vendor products? No Partially, with significant digging Yes, but it is not maintained Yes, maintained as a living inventory with owners

9Have you run any structured AI pilot with defined success criteria? No pilots of any kind Informal experimentation by individual staff One or more pilots without written success criteria One or more pilots with criteria, measurement, and a go/no-go decision

10How much of your back-office document handling (loan files, disputes, member correspondence) is still fully manual? Essentially all of it Mostly manual with some OCR or workflow tooling Key workflows partially automated Major workflows automated with exception handling defined

11Have staff received any AI training beyond a warning email? No Optional self-serve resources shared Structured training for select roles Role-based training program with completion tracking

12If you greenlit an AI project Monday, who would run it? Nobody internal; we would be fully vendor-dependent IT would absorb it on top of existing duties A capable project owner exists but has no AI experience A designated person or small team with prior pilot experience

The credit union AI readiness self-assessment

Scoring model

The questions

Almost there. Where should we send your report?