SpyBara
Go Premium

use-cases/research-to-decision-memo.md 2026-09-28 23:02 UTC to 2026-09-29 22:59 UTC

This page contains 1 addition and 1 deletion.

2026
Fri 18 22:59 Tue 22 23:00 Tue 29 22:59

name: Turn research into a decision memo tagline: Combine evidence, tradeoffs, and open questions into one recommendation. summary: Give ChatGPT research, planning documents, models, dashboards, stakeholder context, and unresolved questions, then ask it to separate evidence from interpretation and draft a sourced decision memo that is ready for leadership review. skills:

  • token: google-drive url: https://github.com/openai/plugins/tree/main/plugins/google-drive description: Read approved recaps, planning docs, models, and source material.

  • token: notion url: https://github.com/openai/plugins/tree/main/plugins/notion description: Check decision history, research notes, and project context when they live in Notion.

  • token: slack url: https://github.com/openai/plugins/tree/main/plugins/slack description: Review stakeholder debate, comments, and unresolved questions from approved channels or threads.

  • token: $spreadsheets description: Validate financial, KPI, or scenario inputs behind the recommendation.

  • token: $documents description: Produce a concise, editable decision memo and review-ready pre-read with source notes. bestFor:

  • Decisions that combine internal evidence, external research, budget, and tradeoffs.

  • Planning or investment questions where the recommendation needs an explicit evidence trail and decision log.

  • Teams that want alternatives, owner confirmations, and unresolved risks visible before a leadership review. starterPrompt: title: Draft a decision memo body: >- I'm deciding whether [team or company] should [decision].

    Use these sources:

    • prior recaps and research: [files or links]

    • planning docs, models, and dashboards: [files or links]

    • audience, account, or market context: [files or links]

    • stakeholder comments and meeting notes: [files, channels, or thread links]

    • budget guardrails: [files or links]

    • decision criteria: [criteria]

    If I ask for web research, keep external findings separate from internal evidence. Write a concise decision memo with a recommendation, supporting evidence, alternatives and tradeoffs, costs, risks, a decision log, open questions, owner confirmations, and source links.

    If the audience needs a pre-read, organize the memo and supporting analysis into a review-ready packet while keeping the memo as the source of truth. Flag assumptions and do not make the decision or share the output. suggestedModel: gpt-6.1-sol suggestedEffort: medium relatedLinks:

  • label: "OpenAI Academy: Everyday work" url: https://openai.com/academy/how-to-use-codex-for-everyday-work/

  • label: "OpenAI Academy: Business operations teams" url: https://openai.com/academy/codex-for-work/how-business-operations-teams-use-codex/

  • label: Plugins url: /codex/plugins

For the complete documentation index, see llms.txt. Markdown versions of documentation pages are available by appending .md to the page URL.

Define the decision before gathering evidence

Decision work becomes clearer when ChatGPT knows the choice, constraints, audience, decision date, and approval boundary. Attach internal context first, then identify any outside research that needs to be completed separately.

  1. State the decision, decision maker, timing, and criteria.
  2. Attach recaps, planning docs, models, dashboards, budget guardrails, stakeholder comments, and meeting notes.
  3. Ask ChatGPT to inventory internal evidence, source gaps, unresolved debate, and requested web research.
  4. Run the starter prompt and review the recommendation, alternatives, tradeoffs, risks, decision log, and missing information.
  5. Verify each material claim and owner confirmation before the memo enters leadership review.

Keep internal evidence, external research, and interpretation in separate sections. If a cost, date, or market fact cannot be confirmed, leave it as an explicit open item. When the audience needs a longer pre-read, keep the decision memo as the source of truth and place supporting analysis in an appendix instead of hiding uncertainty in polished slides.

Pressure-test the recommendation and review packet

Use a follow-up pass to challenge the preferred option, expose which assumptions would change the recommendation, and anticipate the questions decision makers will ask.

Prompt:

Pressure-test the decision memo.

Show:

- the assumptions that drive the recommendation
- the strongest case for each alternative
- evidence that is internal versus external
- costs, risks, and missing information that could change the decision
- claims, numbers, and owner confirmations that still need review
- questions a skeptical decision maker is likely to ask
- the smallest additional analysis needed to reduce uncertainty

Keep the original memo unchanged, cite the relevant source for each finding, and mark every proposed revision for review.