Edwin AI onboarding · Catchpoint evidence

Start with your AI catalog.
Turn critical flows into verified readiness.

First we load the approved agents and their declared dependencies. Then Edwin finds the best agent for a business task. Catchpoint verifies readiness only after the selected critical flow is monitored.

1 · Approved catalog 2 · Business task 3 · Catchpoint evidence

Upload approved catalog

2 · Choose a critical business flow

Ask Edwin which approved agent should handle this task

This step ranks catalog candidates by task relevance, declared trust, and cost policy. It does not claim live readiness—readiness is created by the assurance tests that follow.

No open-web scan Server adapter

1Task

Determine whether an order is eligible for a refund.

2Top agent

Refund assistant asks which approved services can help.

3Discovery

Catalog, registry, MCP, agents.json, CMDB, or gateway inventory.

4Select flow

Confirm the agent and dependencies that make this business journey work.

5Create assurance

Generate Catchpoint tests; verified readiness arrives after they run.

Demo story: Catalog tells Edwin what exists. Relevance chooses what to monitor. Catchpoint supplies the evidence that makes it ready.

1 · Query registry

2 · Read capabilities

3 · Match dependencies

4 · Apply trust + cost policy

5 · Rank relevance

Ranking policy

Relevance selects. Readiness is measured.

1 Discovery finds approved candidates. 2 The owner selects one critical flow and creates its assurance plan. 3 Catchpoint runs the tests; Edwin then publishes verified readiness.

Agent

Not measured Assure this catalog topology

Catalog topology This map stays central: it is built from the selected agent’s declared dependencies. Health journey, dependency, contract, token, and node checks add evidence to this same topology after Catchpoint tests run.

Selected node strategy

Right node, right evidence

Business impact

Economics

Cost root cause

Top dependency issues

Latest 6 hours

Token waste & latency

How we get the data