Facio Blog

Practical notes on human-reviewed AI agents.

Payload-powered product notes, security writing, HITL patterns, and operational guidance from the Facio runtime: long sessions, Placet approvals, audit trails, memory, providers, channels, tools, and Docker-first operations.

Showing 16-20 of 73 articles in Human-in-the-loop.

Human-in-the-loop

HITL and the Second-Order Question: Why the First Action's Outcome Determines Whether the Next Review Is Even Possible

Jul 23, 2026Human-in-the-loop

HITL and the Second-Order Question: Why the First Action's Outcome Determines Whether the Next Review Is Even Possible

Most reviewers think about the action in front of them. Few think about whether the action enables the next review. The first-order question is "is this right?" The second-order question is "does approving this preserve my ability to review the next one?" That second-order question is what separates review systems that scale from review systems that collapse into theater. Here is why the second-order view is the hidden architecture of HITL.

Human-in-the-loop

HITL and the Reversal Question: Why "Can This Be Undone?" Is the Most Important Question the Reviewer Asks

Jul 22, 2026Human-in-the-loop

HITL and the Reversal Question: Why "Can This Be Undone?" Is the Most Important Question the Reviewer Asks

Most reviewers ask "is this right?" before approving. The better question is "can this be undone?" The reversibility question reframes every review — it shifts the focus from correctness to recoverability, from immediate fit to downstream survivability. A reviewer who asks "can this be undone?" first catches more failures, makes better decisions, and produces more durable audit trails. Here is why this one question change makes HITL work.

Human-in-the-loop

HITL and the Pre-Mortem: Why the Reviewer Should Imagine the Failure Before Approving the Action

Jul 21, 2026Human-in-the-loop

HITL and the Pre-Mortem: Why the Reviewer Should Imagine the Failure Before Approving the Action

Every HITL review is a forecast — the reviewer predicts the action's outcome. Most reviewers forecast by asking "will this work?" A small minority ask the inverse: "how will this fail?" That minority is right more often than the majority. The pre-mortem — imagining the failure before approving the action — is the cheapest, most effective single intervention HITL systems can adopt. Here is why, how to design for it, and what it changes.

Human-in-the-loop

HITL and the Judgment Gradient: Why the Same Reviewer Decides Differently on Identical Actions at Different Times

Jul 20, 2026Human-in-the-loop

HITL and the Judgment Gradient: Why the Same Reviewer Decides Differently on Identical Actions at Different Times

Send the same action to the same reviewer at 9am, 11am, 2pm, and 4pm — you will get four different decisions. The action is identical. The reviewer is identical. The decision differs. This is the judgment gradient: the same reviewer applies different judgments at different points in their session. The gradient is the largest source of inconsistency in HITL, and the most invisible. Here is why it happens, what it costs, and how to design for consistency.

Human-in-the-loop

HITL and the Confidence Mismatch: Why the Reviewer's Calibration Rarely Matches the Agent's

Jul 19, 2026Human-in-the-loop

HITL and the Confidence Mismatch: Why the Reviewer's Calibration Rarely Matches the Agent's

Every HITL system has two confidence scores — the agent's and the reviewer's. The scores are supposed to align. They don't. The agent is over-confident on actions that need review and under-confident on actions that don't. The reviewer is over-confident on actions they should reject and under-confident on actions they should approve. The mismatch is the calibration failure nobody measures. Here is why it happens, what it costs, and how to design for alignment.