This guide gives you a reusable starting pattern. It is designed to help you see the work more clearly; it is not a substitute for judgment, source checking, or responsibility for the result.
Prepare the inputs before you ask for output.
The model only sees what you give it. Spend a few minutes naming the reader, desired result, and uncertain information. This makes a first draft easier to assess and reduces the need for decorative rewriting later.
Give the task a useful brief.
Create a decision log from these project notes. Decision question: [state it] Notes and source labels: [paste notes] Known constraints: [time, budget, policy, or technical limits] Return: (1) decision question, (2) options mentioned, (3) evidence or source labels for each option, (4) confirmed decision if one exists, (5) owner and date if stated, and (6) unresolved questions. Do not claim a decision has been made unless the notes explicitly say so. Flag assumptions and missing evidence.
Replace every bracketed field with your real context. Read the output before reuse.
Four steps that keep the result usable.
- 1
Name one decision question rather than a broad project subject.
- 2
Keep source labels beside claims and option descriptions.
- 3
Ask the model to separate confirmed choices from proposals.
- 4
Have the actual decision owner confirm the final record.
A project note is not automatically a decision
Projects collect partial conclusions, proposals, and constraints in the same place. A decision log is useful because it asks which of those points was actually chosen, who owns it, and what evidence supports it.
Keep trade-offs beside the choice
A record that only states the final option can be impossible to revisit later. Preserve alternatives and constraints so a future reader can understand why a choice made sense at the time.
Let the owner close the record
AI can create a clearer draft, but it cannot confirm authority. The person accountable for the decision should review the wording, evidence, and status before the entry is treated as final.
A worked example: choosing a vendor
A team choosing between two email platforms has notes from demos, a security review, and a price sheet. The log should record the decision question — “which platform do we standardize on?” — the options mentioned, the evidence labels for each, and the confirmed choice only if the group actually decided. Keeping the rejected option’s trade-offs in the log is what lets the decision be revisited honestly next quarter.
Entries are for the future reader
The person who will use the log is usually someone who was not in the room. Include the constraint that mattered — budget, policy, deadline — and the owner who confirmed it. A log that only records the chosen option is a summary; a log that records why and who is a usable record.
Run a human check.
- Is the decision question specific enough to answer?
- Are options and evidence tied to the notes that support them?
- Does the record distinguish a confirmed choice from a proposal?
- Has the appropriate owner reviewed the final wording?
AI is strongest here when it makes missing information, structure, and options easier to see. The moment an output becomes a claim, commitment, or decision, bring a person back into the loop.