Inquiry brief

Create a record that a person or assistant can use without reading your previous conversation. This is a book-developed practice artifact. The core exercise needs no code or paid account.

Blank reusable template

Inquiry name, version, and date:

Parent question or attempt:

Project output this inquiry should improve:

Question:

Situation: Who is doing what, under which conditions?

Decision to make:

My answer before assistance:

Competing explanation or alternative:

What I expect to observe:

Evidence that would change my mind:

Available inputs and source locations:

Missing evidence:

Allowed use of the inputs:

Output requested:

Acceptance criteria:

How to handle missing information:

What this inquiry will not establish:

Time, scope, and stopping condition:

Attempt record

What I did, and whether assistance was used:

Input version:

Output location:

What passed the criteria:

What failed, with a specific example:

Revision and its reason:

Fresh example used to check the revision:

What I can explain without the assistant:

Error I can recognize without the assistant:

Unresolved claims and ledger location:

Next question:

Worked sample: fictional Himura Inc. case

Name: inquiry-brief-v1.

Parent: The learner-task branch of curiosity-map-v1.

Output: Revise workshop-brief-v1 while preserving uncertainties in claim-evidence-ledger-v1.

Question: What changes would make this synthetic action-list activity practice the stated learner task and expose unsupported additions?

Situation: Audy is designing a provisional thirty-minute workshop example. Its proposed adult learners would turn meeting notes into action lists. No workshop has been delivered and no customer need has been validated.

Decision: Whether the brief needs a corresponding practice input and checking criteria in addition to a clearer learner outcome.

Starting answer: Audy has provisional criteria from her own earlier inspection. Replacing “understand effective follow-up” with a specific action-list outcome may be enough to guide a fresh draft without carrying those criteria into its instructions.

Alternative: Better wording alone may leave the activity and review criteria unspecified.

Evidence that would change the answer: An agenda can use the revised outcome while never asking a learner to produce an action list; a polished list can pass a vague check while adding unsupported commitments.

Synthetic input

1. The facilitator agreed to circulate the notes by Friday.
2. The group suggested asking the venue about access arrangements; nobody was assigned.
3. The group agreed to try a shorter introduction at the next session; no owner was assigned.

Available evidence: These invented notes, Audy’s draft, and her comparison of input with output.

Missing evidence: Actual educator problems, learner attempts, delivery results, and willingness to pay.

Requested output: Revised practice instructions, a candidate action list, and a check against the notes.

Criteria: Preserve agreed actions; retain only stated owners and deadlines; separate suggestions from commitments; flag missing execution details; require a learner attempt before showing a completed answer.

Missing-information rule: Leave the gap visible and identify a useful clarification. Do not invent an owner, deadline, agreement, interview, or result.

Cannot establish: Learning gains, demand, or viability.

Worked attempt and repair

Faulty candidate: “Facilitator: circulate notes and confirm access by Friday. Organizer: prepare a shorter introduction before the next session.”

Review: The first circulation commitment is supported. The access task was only suggested; its owner and deadline were not supplied. The introduction was agreed, but its owner was invented.

Repair: Keep the circulation commitment. Record the venue proposal as unresolved. Keep the agreed introduction change with the owner explicitly missing. Revise the practice instructions to require these distinctions.

Fresh challenge: “We could send a reminder on Thursday; a decision will be made tomorrow.”

Expected review: Thursday is a proposed timing, not a confirmed deadline. The reminder remains unresolved pending a decision.

Unaided explanation: An agreed action may still lack an owner. A possible action is not made into an agreement by putting it in a tidy list.

Next question: Can a different synthetic input expose a distinction the current criteria fail to cover?

Acceptance criteria

• The brief can be understood without access to the earlier chat.
• The starting hypothesis was saved before the final answer.
• The question points to a decision and names a result that could change it.
• Each claim is supported by an appropriate input or is explicitly provisional.
• The output can be checked against criteria chosen before reviewing it.
• At least one observed defect leads to a targeted repair, or the record explains what meaningful challenge the output survived.
• A fresh example checks the important distinction rather than repeating the same successful output.
• You can explain the decision and identify a plausible error without consulting the assistant.

Common failure and repair

Failure: The question asks for proof of business or learning outcomes using only a synthetic example.

Repair: Narrow the immediate decision to the design property the example can reveal. Name the actual external evidence needed for the larger claim.

Failure: The assistant returns a polished agenda when checking criteria were needed.

Repair: Specify the output directly and provide one acceptable example. Preserve the question if the question itself was clear.

Failure: The prompt asks why the preferred option is best.

Repair: Ask under what conditions each option would be useful, and request a concrete counterexample to the preferred choice.

Failure: You accept the output but cannot explain why.

Repair: Compare one correct and one faulty example, explain the difference, consult the evidence, and try a new pair.

Return to Chapter 4.