wiki/knowledge/sales-enablement/ai-proposal-workflow.md · 767 words · 2026-04-05
AI-Powered Proposal Workflow — Claude + Gamma
Overview
Asymmetric has developed a repeatable, AI-assisted process for generating full marketing proposals in approximately one hour. This replaces a multi-person, multi-day manual process that previously required several rounds of drafting and revision with junior staff.
The workflow requires domain expertise to execute well — knowing what to ask, what to cut, and what the output should look like. AI accelerates the work; judgment still drives the quality.
"Nobody could have just done it. You have to know what you're doing, you have to know what to say, you have to know what not to say, you have to know what images are stupid, you have to know what graphs don't make any sense." — Mark Hope
The Workflow
Step 1 — Research & Data Gathering
Pull competitive and organic intelligence on the prospect:
- Ahrefs — organic keyword data, traffic trends, backlink profile
- SpyFu — PPC history, competitor ad spend, keyword overlap
- Prospect website review — note gaps, messaging, conversion issues
Step 2 — Plan Generation in Claude
Open Claude and provide context in a structured sequence:
- Identify the company, its URL, and business type (B2B, B2C, e-commerce, etc.)
- Paste in the Ahrefs and SpyFu data exports
- Prompt: "This company is interested in marketing. They're clearly looking for leads. Let's put together a plan."
- Review the output — push back on weak ideas, add channel suggestions, refine the strategy
- Iterate until the plan is solid
Example prompt refinements:
- "No, that's not relevant — more of this, less of that."
- "Why aren't you thinking about [channel/tactic]?"
- "What about [specific opportunity]?"
Once the plan is approved, prompt Claude to reformat it for presentation:
"Now put this in a format for Gamma. I want 10 slides or less."
Claude will output a structured slide-by-slide breakdown with all relevant data points, section headers, and copy.
Step 4 — Build in Gamma
- Open Gamma → select Build with AI → choose Prompt input
- Paste the Claude-generated slide outline as the prompt
- Gamma generates the full presentation
- Replace AI-generated stock images with real assets (client screenshots, Asymmetric logo, relevant charts)
- Apply brand colors and logo placement throughout
Typical Proposal Structure
Based on the A3 Environmental example produced in this workflow:
| Slide |
Content |
| 1 |
Executive Summary |
| 2 |
Current State — Website & Online Presence |
| 3 |
Organic Performance & Keyword Analysis |
| 4 |
Competitive Landscape |
| 5 |
PPC Performance (if applicable) |
| 6 |
Key Growth Opportunities |
| 7 |
Recommended 3-Part Growth Strategy |
| 8 |
Implementation & Pricing Tiers |
| 9 |
Why Partner With Us |
| 10 |
Next Steps / Call to Action |
Pricing tiers can be structured by scope — e.g., local ($5,000/mo), regional ($6,500/mo), national ($8,500/mo).
Time Investment
| Phase |
Estimated Time |
| Research (Ahrefs, SpyFu) |
15–20 min |
| Claude plan generation & iteration |
20–25 min |
| Gamma build + image replacement |
15–20 min |
| Total |
~1 hour |
When to Use This
- Any inbound or outbound prospect meeting where a proposal would strengthen the pitch
- Pima leads and referral leads alike — the workflow is fast enough to justify even exploratory calls
- Note: the proposal may not always be appropriate to present on a first call — having it ready is the goal
Tips & Pitfalls
- Don't skip the iteration step in Claude. The first output is a starting point, not a final plan. Push back and refine.
- Replace all AI images. Gamma's default stock images are generic and undermine credibility. Swap them for real screenshots, data visuals, or branded assets.
- Leverage relevant personal experience. If there's a relevant industry background (e.g., Mark's VP of Sales role at Safety-Kleen for environmental clients), include it in the "Why Partner With Us" slide — it differentiates the pitch significantly.
- Tailor the pricing tiers to the prospect's apparent scale and ambition based on what the research reveals.
- [1] — Asymmetric internal operations
- [2] — Inbound lead source feeding this workflow
- [3] — Org structure being built around scaled proposal-to-delivery pipeline
- [4] — Broader Claude use cases across the agency
- [5] — Source meeting where workflow was demonstrated