Can AI draft sales proposals from our discovery call notes?

Yes. AI can turn discovery call notes, the buyer's emails and your approved proposal content into a first draft that restates the customer's problem in their own words and maps your offer to it. The rep edits the draft, and a named person signs off the commercial terms. The AI never invents scope or terms.

#Key takeaways

• The slow part of a proposal is not the template; it is restating the buyer's situation and matching your offer to it.
• A good draft is built from two sources only: what the buyer said, and content your company has already approved.
• Commercial terms, discounts and legal language stay with people, filled in or approved by a named owner.
• Every claim in the draft should trace back to a call note, an email or an approved content block.
• Measure how much the rep rewrites, and how often a proposal goes out with a claim nobody approved.

#Why do proposals take so long to write?

The account executive finishes a good discovery call, promises a proposal by Friday, and then opens last quarter's Word file for a similar customer. The next hours go on find-and-replace, rewriting the problem statement, hunting for the latest case for the integration the buyer asked about, and checking the scope section with the presales engineer.

The result is often a proposal that reads like the previous customer's. Buyers notice. A Gartner, june 2025 survey of B2B buyers found that 73% actively avoid suppliers whose outreach is irrelevant, and 69% reported inconsistencies between a supplier's website and what its sellers told them.

#What can AI draft and what must a person decide?

AI is good at the assembly. It can read the call notes and emails, pull out the buyer's goals, constraints and named stakeholders, and write the situation section in their language. It can then select approved blocks from your content library, such as scope descriptions, delivery approach and security answers, and fit them to what the buyer asked.

People keep three things. The rep decides what to propose and in what order. A named owner, often the sales manager or deal desk, fills in or approves the commercial terms. Legal or security reviews any non-standard clause. The AI leaves those sections marked for a person rather than guessing.

#How do you stop the draft from inventing things?

By limiting where it can take content from. The draft may only use the buyer's own words and blocks your company has approved. Every sentence links back to its source, so the rep can check it in one click. Anything the AI cannot source is flagged, not filled.

This follows advice from model providers themselves. Anthropic, december 2024 recommends starting with the simplest approach that works: a fixed sequence of steps, not an open-ended agent. For proposals that means read notes, select approved blocks, draft, flag gaps, hand to the rep.

#A proposal playbook for your team

1. Collect your last ten winning proposals and mark the sections that were reused nearly unchanged; those become approved blocks.
2. Write a short discovery note template so every call captures goals, constraints, stakeholders and decision process.
3. Keep a list of claims that need approval before they appear in a proposal: integrations, certifications, delivery commitments.
4. Set a rule that commercial terms are filled or approved by one named owner, never drafted.
5. After each proposal, note what the rep had to rewrite and feed it back into the blocks.

#When is a template or a proposal tool enough?

If your proposals are short and your offer rarely changes, a good template in Word or Google Docs plus a ChatGPT team plan for polishing the problem statement may be enough. Proposal software with a content library helps when many people write proposals and the library is well kept.

A custom workflow is worth it when proposals are long, scope varies by customer, notes and emails sit across the CRM and inboxes, and buyers write in Arabic and English. For formal questionnaires, see our separate post on rfp responses from an answer library.

#Where does Stennir fit?

Stennir starts with an AI consulting audit that ranks proposal drafting against your other sales workflows, with a cost, a payback case and a hosting region. The report is yours whether or not you build with Stennir. If proposals rank first, AI implementation builds the drafting step into your CRM and document tools, with the rep editing and a named owner approving terms, every action logged. You pay for finished work.

#What should a sales leader do next?

Time your next three proposals from call to send, and mark which sections were rewritten from scratch. That shows where drafting helps. See AI for sales, then take the free readiness assessment or book a 30-minute call.

#Common questions