A technical project manager launched an agentic workflow that is mid-run migrating 300 help-center articles, and stakeholders just clarified that API examples must stay in Python. The run cannot restart without losing an hour. Which interaction should the manager use to interact with an agentic workflow while it is running?
Select an answer to reveal the explanation.
Short Explanation
You can still nudge the wheel while the car is moving. A mid-run steering note redirects what’s left and live steering saves your hour while keeping the new Python rule front and center.
Full Explanation
The correct interaction is in-flight steering because agentic workflows support human-in-the-loop guidance that adjusts behavior for future steps without discarding checkpointed progress, which directly satisfies the dual constraints of enforcing Python for work not yet done while avoiding a costly restart. One distractor fails on scope precision because improving formatting and clarity mid-run is a legitimate steering use but it leaves the newly clarified language constraint unaddressed, so stakeholders would still receive non-Python examples. Another distractor fails on state management because demanding immediate reprocessing of completed articles confuses mutable future trajectory with immutable completed work, forcing redundant work, risking inconsistency and undermining the explicit goal of preserving the hour already invested. The remaining distractor fails on continuity because pausing to rewrite the base system prompt and restarting is a sound persistent governance practice for the next execution but it violates the no-restart condition and sacrifices in-progress state for a problem that steering can solve live. Exam caveat: distinguish live steering for remaining steps from prompt revision for future runs and from bulk remediation of finished outputs, as wording about mid-run, remaining, and without restart signals the expected pattern. Practical check: after sending the steering message, verify the workflow acknowledgment and inspect the next few migrated articles to confirm Python-only examples before allowing the full batch to continue.