For us, coverage means the important user journeys are covered and stay covered. We start with likely flows from the codebase, then let users point the agent at the journeys that matter most. We also already have session replay agents in production to find bugs, and we’re now using that same system to spot coverage gaps and generate tests for missing flows.