01How to use it
Scoring rule
The thresholds are Ace8's working rule, not an industry standard. Adjust them for your risk level.
- 01
Answer yes or no
Yes if you can show evidence today. A plan to do it is a no.
- 02
Score each area
One point per yes. Each area scores 0 to 4. Total is out of 20.
- 03
Find the floor
The lowest area score matters more than the total. One weak area stalls the rest.
- 04
Read the result
Total 16 to 20 with no area below 3: ready to scale. Total 10 to 15, or one area at 2: ready for a bounded pilot. Any area at 0 or 1, or total under 10: fix foundations first.
02Area 1: Data
Data checks
- You know where the data you would use lives, and who owns each source.
- You can say which data is sensitive or regulated, and it is labeled as such.
- Data quality is measured, with known gaps written down.
- You control what leaves your network, and can show what has been sent to an outside AI service.
03Area 2: Platform
Platform checks
- You have a build and test pipeline that runs on every change.
- You have non-production environments where AI-assisted changes can be tried safely.
- You can approve or block specific AI tools and models centrally.
- You can monitor cost and usage of any AI service in use.
04Area 3: Code
Code checks
This area decides whether AI-assisted modernization is safe. Without tests, a reviewer has nothing to check an AI change against.
- Your key applications have automated tests you trust, and you know the coverage.
- You have an inventory of applications with runtime and framework versions.
- Dependencies are tracked, and known vulnerabilities are visible.
- Platform-specific code, such as vendor APIs and descriptors, is identified.
05Area 4: People
People checks
- Named engineers own the review of AI-written changes.
- Reviewers have been shown the typical failure modes of AI output.
- There is time budgeted for review, not just for generation.
- Someone senior is accountable for the outcome of the AI program.
06Area 5: Governance
Governance checks
- A written policy says what AI tools may be used, on what code and data.
- Licensing and intellectual property questions about generated code have been answered by legal.
- Changes made with AI are recorded: tool, version, who approved.
- There is a rollback path for every AI-assisted change.
07After the score
What to do with a weak area
This list is a self-check, not an audit. A sample scored assessment report is being prepared and will be linked here.
| Weak area | First fix |
|---|---|
| Data | Classify sources and block sensitive data from external services |
| Platform | Add a test pipeline and a sandbox environment |
| Code | Inventory applications, then write characterization tests on the riskiest ones |
| People | Name reviewers and set review time |
| Governance | Write the one-page tool policy and the change log |
08Next step
How Ace8 helps
Ace8 is the AI-assisted application modernization practice of AceMQ. The AI readiness assessment turns this checklist into evidence: we check each line against your code, pipeline and policies, then give a ranked list of fixes. If your code score is the weak one, the modernization assessment is the better start.