Question 1

How is 'AI technical debt' actually measured here?

Accepted Answer

Ten factors: how many AI systems, models, agents, and prompts you have and how well they're tracked, how many integrations and how much duplicated infrastructure exist, how much manual work still props things up, and how much of it has documentation, evaluation, and monitoring.

Question 2

Is having multiple AI systems automatically technical debt?

Accepted Answer

No. Running many AI systems isn't the problem — not tracking, documenting, or governing them is. A company with 15 well-documented, monitored systems scores lower debt than one with 3 nobody can fully account for.

Question 3

What's the difference between this and the AI Production Readiness Assessment?

Accepted Answer

The Production Readiness Assessment scores whether one specific AI system is engineered to hold up in production. This calculator scores accumulated debt across your entire AI portfolio — inventory, duplication, and governance, not the engineering quality of any single system.

Question 4

What counts as a 'good' score?

Accepted Answer

Low Debt (0-24) is the goal, but most organizations land in Moderate to High — fragmented AI systems with inconsistent documentation and evaluation is the common state, not an unusual one.

Question 5

What happens after I see my score?

Accepted Answer

Your highest-debt categories map to specific remediation actions. When debt spans multiple systems, the fix is usually structural — a shared AI platform or a governance layer — rather than fixing each system individually. That's what AI Platform Engineering and AI Governance & Control cover.

AI Technical Debt Calculator

How many distinct AI systems are running, and how well are they tracked?

About the Assessment