The Agentic Readiness Index scores how prepared an institution is for AI agents that act on behalf of a user, not just answer questions about it. It measures six dimensions, each on a 0 to 100 scale, and rolls them into a single readiness score.
Visibility readiness asks whether an AI can find and describe you. Agentic readiness asks something harder: whether an AI agent could actually use you to get something done. Most financial institutions are not measuring the second question yet.
The six dimensions
- API completeness. Whether an agent can actually do things with you through documented, working interfaces, not just read about you.
- Policy readability. Whether your rules, terms, and constraints are written so a machine can parse and respect them.
- Identity interoperability. Whether your identity and data line up with the standards agents use to recognize and verify you.
- Agent directive clarity. Whether you give clear instructions for how agents should and should not use your site.
- Audit and recovery. Whether agent actions can be traced, and mistakes caught and reversed.
- Human handoff quality. Whether an agent can hand a task back to a person cleanly when it should.
How the scoring bands work
Each dimension is scored from 0 to 100, then combined into an overall Agentic Readiness score with three bands:
- 80 to 100, Leading. An agent can find you, understand your rules, and complete real tasks with confidence.
- 50 to 79, Building. The foundation is there, but gaps in policy, identity, or handoff would trip an agent up.
- Below 50, At risk. Agents cannot reliably act with you, and that gap will show the moment your competitors close theirs.
What the full method includes
The bands above are the public version. The full rubric, the per-dimension checks, the weighting, and the remediation plan, is the work I do with clients through Atlas Instinct. If you want your institution scored against it, that is where to start.
Get scored through Atlas Instinct, or email kevin@kevinfarley.org.