The Decision Memo template makes your AI work a decision through four steps: frame it with explicit criteria, lay out the realistic options even-handedly, write a recommendation tied to those criteria, then record the call. The recommendation stops for you, because the decision is yours. Reject it and the AI reconsiders the options.
4 steps · 1 approval · 1 loop back
example · the steps are this template's real steps
For founders, managers and operators facing build-or-buy, vendor or hiring calls they will need to explain months later.
It is a stored workflow, not a prompt. Your AI client receives one step at a time from ConvOps over MCP, does the work, and advances. The engine holds the run at the approval until a person says yes.
4 steps, in this order. Below each one is the instruction your AI receives at that step, exactly as the template stores it.
Frame the decision before exploring options.
Capture:
A decision with no stated criteria cannot be evaluated. If criteria are unclear, agree them with the operator before advancing.
Save the framing as a task note with task_notes_add. The next step (Options) reads this note.
Read the task notes with task_notes_list: the note from the Frame step contains the decision, stakes, and criteria. Work from it.
Lay out the realistic options. What good looks like:
Be even-handed: do not stack the deck toward a favorite. If a research report informs this, pull from its findings.
Save the options and their trade-offs as a task note with task_notes_add. The Recommendation step reads this note.
Read the task notes with task_notes_list: the Options note holds the options and trade-offs, and the Frame note holds the criteria. Work from them.
Write the recommendation:
Present this as a recommendation, not a verdict. Save the recommendation as a task note with task_notes_add. Then ask the operator to make the call. This is a checkpoint: the decision is the operator's to make, so the AI cannot advance until you decide. If you reject the recommendation, the workflow returns to Options to reconsider.
rejected → back to Options
Read the task notes with task_notes_list: the Recommendation note holds the recommended option and the operator's decision. Work from it.
Record the final decision. Save a task note with task_notes_add capturing:
This note is the durable record of the decision. Close out the task.
One step waits for a person. You answer in the chat; the engine records the verdict and the audit row records who gave it.
The AI recommends; you decide. Approve to record the call, or reject and the workflow returns to Options to reconsider.
what you review
reject → back to Options
gates the engine checks without you
Read from the template's own steps: the tools they call, the files they write, and whether they touch code.
Connect your AI client first, then install Decision Memo from the catalog in the ConvOps app. See how to connect a client
An installed template is a workflow in your workspace. Ask your AI to change it in plain words; the edit lands on the one step that needs it.
new chat
more example prompts
The decision made, the date and who made the call, the rationale (the criteria it satisfied and the trade-off accepted), and the options not chosen, saved as a task note that is the durable record.
Yes, where relevant. The Options step names each genuine option including "do nothing", scores each against the criteria, and is told not to stack the deck toward a favorite.
Yes. The Research Report template's final note is its named input for a Decision Memo, and the Options step pulls from a research report's findings when one exists.
Any MCP client connected to ConvOps. Verified clients: Claude Code, Claude (desktop, web, Cowork), OpenAI Codex, ChatGPT, Cursor, OpenCode. It needs no repository, so a plain chat client works.