Eight questions, at the level you actually run the box. Not "do you have a strategy" but is journaling on for the files that matter, can an agent reach Db2, who is allowed to elevate. The platform gives you more than you are told. This reads how much of it you have claimed.
Pick the description that fits, then nudge the number. Nothing leaves your browser.
What the platform enforces at the microcode boundary, where IBM i is structurally strong. The question is how much of it you have kept.
Whether an agent can reach in, and whether autonomy is staged. Where the real work usually is.
You own the governance a cloud shop is still trying to bolt on: object security, journaling, an audit journal, adopted authority. The platform will let an agent act safely. What it cannot do is reach in. Your exposure is the modernization surface: an agent has to be able to get to the logic and the data, and most of that is still behind 5250. That is the work, and it is a very different message than "you are behind." You are not behind on safety. You are closed on access.
Rare air: you have kept the platform's governance and opened a surface an agent can reach. The remaining question is not technical readiness. It is whether the organization will authorize the autonomy your platform can already carry safely. That is the gate that halts next, and it is not on the box.
IBM i hands you agent-grade governance for free, and the scores say you have not claimed much of it. *ALLOBJ where you meant least-privilege, journaling off where you meant reversible, QAUDJRN unread. The good news is that this is claiming, not building: the mechanisms are already under you. Turn them on before you point an agent at anything.
You have modernized the surface, so an agent can reach in, but the safety boundary is thinner than the access. That is the dangerous order: reachable before it is governed. Close the boundary (identity, reversibility, attribution) before the reach becomes a liability.
This read measures the platform. The one thing it cannot score, and the one most likely to gate you anyway, is whether you have the people to carry it: the RPG generation that holds the logic is retiring, and agent-readiness leans hard on who is left to modernize and govern. That is a people question, and it deserves its own read.
This read is adapted from the Pegasus4i Agent-Ready Read. If you want to see how the platform enforces each of these, and where the gate stops being a platform question, the mechanism is mapped on pegasus4i ↗
The four safety areas map to the four hard-stop questions of the ExO 3.0 CIO readiness diagnostic (Appendix F), used with permission. What ExO 3.0 is.