Workshop-brief practice specimens

Use with the capability baseline in Chapter 1 and claim-evidence ledger in Chapter 2. Every input and output below is an authored fictional teaching specimen. None is a transcript of an actual AI run, an interview, or evidence of customer demand. The set makes selected example verdicts inspectable; it does not substitute for running and retaining your own five attempts.

Initial input

Project: Himura Inc., a proposed research and learning-design studio.
Audy has experience facilitating workshops. She is considering a short
workshop about meeting follow-up for adults. She has not interviewed the
intended learners or tested demand. The topic is provisional.

Draft a short design brief from these notes. Identify missing information.

Constructed flawed response

Audience: Busy professionals who lack confidence in meeting follow-up
and prefer hands-on activities.

Outcome: Participants will understand effective follow-up.

Design: Use group discussion and an interactive worksheet, which will
make learners more confident and ensure that they apply the learning.

Annotated review

The input does not establish low confidence or an activity preference. Those statements must be removed or presented as hypotheses to investigate. “Understand effective follow-up” does not specify a performance the reviewer can inspect. The final sentence promises learner outcomes that the input and this drafting exercise cannot establish.

A reviewer can identify these problems by comparing the answer with the supplied text. They do not need to know whether some real professionals lack confidence; the issue is that the answer presents an unsupported assertion as a fact about this proposed audience.

Constructed correction

Established by the notes: Audy has facilitation experience and is
considering a short workshop on meeting follow-up for adults.

Unresolved: The specific audience, learner task, current difficulties,
activity preferences, workshop duration, and demand for the offer.

Proposed next step: Ask what a participant should be able to produce or
do after the workshop, using an example of the work they encounter.

Question: What concrete output should the participant create, and what
input would they use to create it?

The correction does not complete the workshop design. It preserves the gap that has to be resolved before a useful design can be specified.

Fourth input: missing-task variant

This specimen corresponds to the missing-information trial described in row 4 of the baseline example.

Audy is considering a short workshop about meeting follow-up for adults.
She wants a provisional brief, using only the information supplied here.
No learner task, output, or performance criterion has been specified.

An acceptable response identifies the missing task and asks a focused question, such as: “What should a participant be able to produce from a set of meeting notes?” It must not silently decide that the workshop has already selected that outcome.

The observation supporting claim C02 is narrow and inspectable: this supplied input explicitly leaves the learner task unspecified.

Fifth input and corrected output

This specimen corresponds to the ordinary-input trial described in row 5. It supplies additional information so the response can propose an inspectable output. The workshop remains hypothetical.

Audy proposes a practice workshop in which adults use synthetic meeting
notes to create an action list. She wants the brief to distinguish agreed
actions from suggestions. Owners and deadlines should remain unknown when
the notes do not supply them. No real learner trial or customer interview
has taken place. The offer and audience remain provisional.

A constructed acceptable output is:

Provisional learner task: Create an action list from supplied synthetic
meeting notes, preserving the difference between agreements and suggestions.

Inspectable output: An action list in which each agreed action points to
the supporting note. Unspecified owners and deadlines are marked unknown.
Suggestions are retained separately as unresolved possibilities.

Current limits: This is a proposed practice design. Learner improvement,
audience needs, and demand have not been tested.

Next design question: Which example notes would reveal an invented owner,
deadline, or commitment when the list is reviewed?

The answer retains the input’s uncertainty and proposes something the reviewer can inspect. These properties support C04, limited to this constructed output and these requirements. They do not establish that any real model reliably produces it. A reader’s own C04 record should point to their actual saved output, prompt version, input, and review.

Use the specimens actively

Before reading the annotated review, mark every sentence in the flawed response that goes beyond the input. Then write your own correction. Compare it with the supplied one and explain any difference you want to keep. Later, run your saved prompt with your own synthetic inputs and retain the results. Keep constructed examples, actual runs, and observed customer evidence visibly separate.