Operators are moving quickly on agentic AI, yet most of what is labelled agentic today still leaves the final call with a person. This Appledore Research Note examines whether the data behind those agents is ready for them to find, understand and safely act on.

We put two questions to network, data, architecture and AI leaders at operator groups in Asia, Europe and North America: is your data ready for AI agents, and what are you doing to make it so? We set out the deficiencies holding operators back and the seven principles they are converging on, and we argue that much of the challenge lies in how data is owned, linked and governed rather than in the AI itself. We also show where operators still differ, on sequencing, proving periods and how they build context.

This note is for CSP network, data and AI strategy teams, and for the OSS, observability and data platform suppliers that serve them. Readers get a practical framework for agent-ready data and concrete recommendations for both operators and vendors.

Companies mentioned in this report: Amdocs.