AIYING · V4.0 · AI LAB
Make a Brief Generate Its Own Missing Questions
See differently. Generate possibilities. Accomplish what matters.
By Ravindra Dastikop
A client sends a brief:
“We need an introductory AI workshop for 40 employees next Friday. Make it practical. Participants will use their own laptops. Please send a proposal tomorrow.”
You could ask AI to write the proposal immediately.
In this experiment, you will build something that improves the brief before the proposal is written: a question sequence in which each answer helps determine what to ask next.
The accomplishment is a reusable workshop intake conversation, tested against a fictional brief.
The hidden assumption
We often treat a brief as the starting point for producing a document. But a brief also contains clues about what we do not yet know.
“Practical” could mean following a demonstration, completing an exercise, or producing something useful for work. Those interpretations require different workshop designs.
Your experiment
- Tool: Any text-based AI assistant you already use.
- Materials: The fictional brief above.
- Suggested time: 20–30 minutes.
- Expected output: A tested intake conversation and a revised brief.
Keep the fictional information throughout the experiment so you can share and repeat it easily. This is a lab protocol; the outcomes below are checks to perform, not claims of a completed experiment.
Step 1 — Separate what is stated from what is missing
Paste the brief and this prompt:
Read this workshop brief. Do not write a proposal yet.
Create a table with four columns:Keep facts separate from assumptions. Do not invent answers. Prioritize questions whose answers could change the workshop design.
- Information explicitly stated.
- What remains unknown.
- A decision that depends on that missing information.
- One question that would help resolve it.
Check the result yourself. The brief states 40 employees, next Friday, own laptops, and proposal tomorrow. It does not establish duration, participant experience, permitted tools, internet access, or the intended work output.
An assistant that fills these gaps has made assumptions. Identify them before proceeding.
Step 2 — Make each answer influence the next question
Continue with this prompt:
Read this workshop brief and conduct an intake conversation.
Ask one question at a time. After each answer, briefly state its design implication and select the next most useful question.
Keep track of facts, client decisions, recommendations, and unresolved items separately. Do not treat unanswered questions as agreed decisions.
When enough information is available for a draft brief, ask whether I want to generate it.
Now answer the questions as a fictional client. For example:
“Participants should produce a first draft of a client follow-up email.”
That answer should lead to questions about the email task, suitable practice material, and how participants will judge the draft.
It should influence the conversation.
Step 3 — Test whether the answers change the route
Run the intake prompt in two fresh chats, each with the same original brief. At the first question about tool access, provide the relevant answer below. Keep other answers consistent where possible.
| Test | Fictional client answer | Expected consequence |
|---|---|---|
| A | Participants may use an approved AI assistant. | Investigate access, accounts, and suitable practice inputs. |
| B | Participants cannot access AI tools during the session. | Clarify how the intended hands-on accomplishment can be achieved under that constraint. |
Pass condition: The next questions respond to the different constraints.
Fail condition: Both conversations continue through substantially the same checklist without addressing the difference.
Do not judge the conversation only by how fluent it sounds. Look for a visible change in direction.
Step 4 — Introduce a contradiction
Give the assistant these two answers during one test:
“Every participant must complete the exercise individually.”
“Only five laptops will be available for the 40 participants.”
Does it notice the conflict? The second answer also conflicts with the original expectation that participants will use their own laptops.
A useful response should identify the tension and ask which requirement or arrangement can change.
It might offer group work or rotating access as possibilities, but it should seek your decision before recording either as agreed.
Step 5 — Produce the revised brief
Finish with:
Produce a revised workshop brief using only the information established in our conversation.
Include:Then show which answer supports each major design choice. Identify any recommendations you have added as recommendations. Preserve unresolved contradictions explicitly.
- Intended participant accomplishment.
- Audience and starting experience.
- Duration and delivery conditions.
- Tool access and practice materials.
- Evidence that the exercise succeeded.
- Unresolved decisions.
Compare this with the original brief.
Has “make it practical” become something you could design, deliver, and check?
Record what actually happened
Copy and complete this experiment log. Record the tool or model label displayed, the date, your prompts, and the relevant answers so another attempt can be compared.
| Check | Your observation |
|---|---|
| Kept stated facts separate from assumptions | |
| Asked a useful clarification of “practical” | |
| Changed questions when tool access changed | |
| Detected the laptop contradiction | |
| Preserved unresolved decisions in the final brief |
These are checks to perform, not results to assume. If the conversation fails, record the failure and revise the prompt.
A successful run demonstrates this particular test; it does not guarantee the same behavior on every future brief.
Where the situation becomes generative
The useful shift occurs when an answer changes what the conversation investigates next.
Question → answer → design implication → next question.
AI helps conduct that cycle. The client supplies context and decisions. You judge whether the resulting brief is sufficient.
Accomplish what matters
Your completed experiment should leave you with three things:
- A reusable intake prompt.
- A revised fictional brief.
- A test record showing where the conversation adapted or failed.

