The Proposal template makes your AI build a client-ready proposal: capture the requirements, define scope and pricing, draft the proposal with those figures unchanged, then check every number and deliverable against what you approved. It stops twice: on the scope and price, and before the proposal goes to the client.
5 steps · 2 approvals · 1 loop back
example · the steps are this template's real steps
For agencies, consultants and sales teams who quote services and cannot afford a wrong figure in a sent proposal.
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 each of its two approvals until a person says yes.
5 steps, in this order. Below each one is the instruction your AI receives at that step, exactly as the template stores it.
Capture the client's requirements before scoping anything. Do not price yet.
Record:
If this comes from a discovery call, pull from those notes. If requirements are ambiguous, list the questions for the operator rather than guessing.
Save the requirements as a task note with task_notes_add. The next step (Scope & Pricing) reads this note.
Read the task notes with task_notes_list: the note from the Requirements step contains the client's needs, deliverables, and constraints. Work from it.
Define scope and pricing. What good looks like:
Save the scope and pricing as a task note with task_notes_add. Then ask the operator to approve the scope and the numbers before any drafting. This is a checkpoint: committing to a price is the judgment moment, so the AI cannot advance until you approve.
no rejection path · the run moves on only when approved
Read the task notes with task_notes_list: the Scope & Pricing note holds the approved numbers and deliverables, and the Requirements note holds the client's context. Work from them.
Write the full proposal:
Save the full proposal as a task note with task_notes_add. If it exceeds 10,000 characters, split across notes or store where the operator keeps documents and record the location. The Client Review step reads this note.
Read the task notes with task_notes_list: the Draft note holds the proposal and the Scope & Pricing note holds what was approved. Work from them.
Check the proposal before it goes to the client:
Ask the operator to approve the proposal for sending to the client. This is a checkpoint: sending to the client is irreversible, so the AI cannot advance until you approve. If you reject, the workflow returns to Draft.
rejected → back to Draft
The proposal is approved and ready to send to the client. Confirm the final proposal is saved as a task note, then close out the task.
Two steps wait for a person. You answer in the chat; the engine records the verdict and the audit row records who gave it.
Committing to a price is the judgment moment. You approve the deliverables, timeline, pricing model, exclusions and assumptions before any drafting.
what you review
no rejection path · approve to continue
Sending to the client is irreversible. Every figure must match the approved pricing exactly. Reject it and the run returns to Draft.
what you review
reject → back to Draft
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 Proposal 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
No. The Draft step writes scope, timeline and pricing exactly as approved, and Client-Ready Review checks every figure matches the approved pricing and no deliverable was added or dropped.
Yes. Requirements pulls from discovery notes when they exist and lists open questions for you rather than guessing. A Meeting Follow-Up run often precedes it.
Each deliverable with what is and is not included, a timeline or phases, pricing per deliverable or a total with the model (fixed, retainer, hourly), explicit exclusions and the assumptions the price depends on.
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.