# Data and Capability Readiness — AI Strategy, Ethics and Governance

Source: https://www.geekswithgeeks.com/en/ai-strategy-governance/strat-readiness

> Assess whether data, infrastructure, skills and processes can support a use case.

## Four things to check

**Data**: is it available, accurate, up to date, legally usable and free of unnecessary personal information? **Technology**: can it connect to your systems securely, and where will it run? **People**: do you have product owners, engineers, domain experts and reviewers with time for this? **Process**: how will outputs be reviewed, who handles exceptions, and how will the work change for users? Many projects stall on data access and ownership, not on the model, so check these early.

## A readiness checklist

Answer honestly. Two or more "no" answers mean fix the gap before the pilot, not after.

```text
Data      [ ] we know where it lives and who owns it
          [ ] it is accurate and current enough to rely on
          [ ] we may legally use it for this purpose
Tech      [ ] a secure connection to the source systems exists
          [ ] we know where the model runs and what data leaves our network
People    [ ] a named product owner and a domain expert have time
Process   [ ] reviewers are assigned; exceptions have an owner
Measure   [ ] a baseline of today's metric exists to compare against
```

## Measure the baseline first

You cannot show improvement without knowing today's numbers. Record current handle time, error rate or cost before the pilot so the result is credible.

**Quiz:** Why record a baseline before a pilot?

- [ ] Baselines are decorative
- [x] Without it you cannot show real improvement
- [ ] It speeds up the model
- [ ] Regulators ban pilots otherwise

*Answer:* Without it you cannot show real improvement. Comparing against a known starting point is what makes the result trustworthy.