Make the handoffs visible

1846. Lesson identity

Module

4.12 — Teams and roles

Lesson

Make the handoffs visible

Academic type

Workflow

Estimated time

About 50 minutes, including the practical assessment and knowledge check

This lesson turns informal ownership into a visible working agreement. You will identify who coordinates work, who accepts it, what evidence supports acceptance, what the decision excludes, and what happens when the handoff is blocked.

1847. Learning objective

After this lesson, you can create a responsibility matrix for a small game team that assigns explicit human ownership, separates contribution from accountability, and defines observable acceptance evidence, decision boundaries, and actionable escalation paths.

1848. Why this matters

Work can disappear between roles when a task has contributors but no clear owner, when two people interpret “done” differently, or when nobody knows who decides after a disagreement. AI-assisted work can increase this ambiguity because an AI tool may draft or compare work without being able to accept its consequences. A responsibility matrix makes each decision boundary visible while keeping a named human accountable.

1849. Prior knowledge

You should have completed 4.12, Lesson 1: AI can contribute, but it is not accountable. You should be able to distinguish a contribution from a human approval decision.

1850. Core concept

A handoff is complete only when the agreed evidence has been checked and a named human accepts or rejects the result.

Each matrix row records:

1. Responsible: the human who performs or coordinates the work.
2. Accountable approver: the one named human who accepts or rejects it.
3. Contributors: people or tools that provide input, including AI when appropriate.
4. Acceptance evidence: the observable result used to judge the work.
5. Boundary: what this approval does not decide or authorize.
6. Escalation trigger and path: the condition that pauses the handoff, the next decision-maker, and the action required.

One person may be both responsible and accountable in a small team, but that combination should be deliberate. AI may be listed as a contributor, never as the accountable approver.

1851. Mental model: a VISIBLE handoff

Use these questions to inspect a row:

1852. Model row

The first row below is complete. Notice that its acceptance evidence describes an observable result, its boundary prevents the approval from expanding silently, and its escalation names both a trigger and a decision-maker.

The programmer and AI contribute, but neither automatically owns the player-facing approval. The evidence is stronger than “reviewed” because another person can inspect it.

1853. Guided practice

Complete the two partially prepared rows. Spend about five minutes on each.

For each guided row, ask:

• Can the accountable approver make a clear accept-or-reject decision from the evidence?
• Does the boundary prevent that decision from being mistaken for a broader approval?
• Does escalation identify a trigger, a human decision-maker, and an immediate action?

Revise weak cells directly rather than writing separate rejection questions.

1854. Independent practice: complete the eight-area matrix

Copy the model and guided rows into one matrix, then complete the five remaining rows. The area labels and review prompts are already supplied so that you can focus on the decisions.

Use fictional names if you are not working with a real team. Do not use “the team,” “someone,” or “AI” where a specific human decision-maker is required.

1855. Focused review

Review the matrix once, in place:

1. Correct any row whose evidence is only an activity such as “reviewed” or “tested.” Replace it with an observable result.
2. Correct any escalation cell that lacks a trigger, named human decision-maker, or immediate action.
3. Highlight rows where the same person executes and approves the work, or where a responsibility has no credible owner.
4. For at least one highlighted small-team or solo-project gap, add a compensating control such as a second-person review, an evidence-based checklist, a scheduled escalation, or a hold point before acceptance.
5. Add a brief rationale only to the two handoffs you judge highest-risk. Explain why their ownership, boundary, evidence, or escalation design reduces ambiguity.

1856. AI-assisted workflow

AI may help draft role descriptions, challenge vague evidence, or identify missing handoffs. Give it the proposed matrix and ask which cells still permit conflicting interpretations. Inspect every suggestion yourself. A human team member must decide assignments, boundaries, sufficiency of evidence, and approval outcomes.

1857. Practical assessment and evidence

Submit the completed matrix and the two short high-risk rationales for the planned practical assessment Visible handoff responsibility matrix. The assessment checks:

• explicit human ownership;
• separation of contribution from accountability;
• observable acceptance evidence;
• clear decision boundaries;
• actionable escalation;
• coverage of all eight areas;
• treatment of at least one small-team or solo-project ownership gap.

The quiz that follows is a knowledge check. The matrix is the performance evidence for the lesson capability.

1858. Key takeaways

• A handoff finishes when evidence is checked and a named human accepts or rejects the result.
• Responsible, accountable, contributor, evidence, boundary, and escalation fields answer different coordination questions.
• AI can contribute to a handoff but cannot own its approval.
• Small teams need explicit controls for responsibilities that would otherwise disappear.

1859. Next lesson

Continue to 4.13 — Outsourcing. An outsourcing request is a cross-boundary handoff, so its owner, acceptance evidence, exclusions, and escalation path must be explicit before another party can act on it.

1860. Knowledge check