What this automation does
After a discovery or sales call, this automation takes the call transcript, extracts the prospect's pain points, requirements, budget signals, and timeline. It then maps those requirements to your product or service offerings and generates a draft proposal document — complete with executive summary, recommended solution, pricing, and timeline.
Writing proposals is one of the most time-consuming parts of the sales process. Reps spend 2-4 hours on each one, often copying from previous proposals and adjusting. This automation produces a solid first draft in under 5 minutes that captures the specific needs discussed on the call.
Tools you need
- Fireflies.ai or meeting transcription tool: Provides the call transcript as input
- OpenAI API: GPT-4 analyzes the transcript and generates the proposal ($0.15-0.30 per proposal)
- Google Docs API or PandaDoc: Creates the formatted proposal document from AI output
How to set it up
Step 1: Create a proposal template in Google Docs or PandaDoc with sections for: executive summary, prospect's challenges (pulled from the call), recommended solution, scope of work, pricing, timeline, and next steps. Mark each section with placeholder tags.
Step 2: When a new transcript is available, trigger a Make scenario. Send the transcript to OpenAI with a detailed prompt that includes your product catalog, pricing tiers, and typical project scopes. Ask the AI to extract: prospect company and contact info, stated pain points and goals, budget range mentioned, decision timeline, and which of your offerings best address their needs.
Step 3: Take the AI's structured output and populate your proposal template. Use the Google Docs API to replace placeholders with the generated content. For PandaDoc, use their API to create a new document from your template with pre-filled fields.
Step 4: Send the draft proposal to the account executive for review. They typically spend 15-20 minutes refining rather than 2-4 hours creating from scratch. Track win rates for AI-assisted proposals versus manually created ones to measure impact.
Cost breakdown
| Item | Cost | Notes |
|---|---|---|
| Transcription tool | $10-$20/mo | If not already subscribed |
| OpenAI API | $15-$30/mo | ~$0.20 per proposal at 75-150 proposals/mo |
| Make or Zapier | $15-$25/mo | Based on proposal volume |
| PandaDoc (optional) | $0-$20/mo | Free tier available |
| Setup time | 45-90 min | One-time |
| Total monthly | $40-$95/mo | Saves 6+ hours/week of proposal writing |
Frequently asked questions
AI handles 80% of the proposal — structure, standard language, requirement mapping, and pricing. The remaining 20% — strategic positioning, custom terms, and relationship nuances — still needs human refinement. But starting from a solid draft instead of a blank page saves significant time even for complex deals.
The AI works with whatever information is available. If the transcript is poor, supplement it with CRM notes or a brief form the rep fills out after the call — company size, key pain points, budget range. The proposal quality directly correlates with input quality.