AI SERVICE Proposal-to-Close Engine

One command. Five writing agents. One proposal that closes.

Your pitch got the meeting — your proposal is what loses the deal. This engine researches the prospect, builds 3-tier ROI-anchored pricing, phases the scope, quantifies the value, and pre-handles objections — then scores the document 0–100 before you send it.

research 20% pricing 20% scope 20% value 20% positioning 20%
0–100 scored,
before you hit send
5
Specialist agents draft the proposal in parallel
0100
Proposal Strength Score — a go/no-send gate
14
/proposal commands: build, pricing, ROI, SOW & more
4
Follow-up emails generated with every proposal (Day 0/2/5/9)
Try the engine — live

Watch a C-grade proposal become an A+.

Below is a proposal document for a fictional Dubai design studio, in the exact structure the engine writes. Flip each agent on and its section is rebuilt in place — while the real 5-category scoring rubric recomputes the Proposal Strength Score live.

Build the proposal

Each switch is one of the engine's five parallel agents — each carries 20% of the composite score.

Weak — likely to lose, rework needed.
Gate closed — don't send

Proposal Strength Score 42 out of 100, Grade C

70–100 40–69 0–39
PROPOSAL-Saffron-Slate-Interiors.md Capability demo · not a client case study
PROPOSAL

AI Search Visibility Strategy for Saffron & Slate Interiors

Prepared for: Saffron & Slate Interiors (fictional) Prepared by: [YOUR AGENCY] Valid: 30 days
42/100 Grade C Weak — likely to lose, rework needed

Executive Summary

We are pleased to submit this proposal for digital marketing services. Our team has many years of experience delivering high-quality solutions for businesses of all sizes, and we are confident we can help Saffron & Slate grow its online presence. We look forward to the opportunity to work together.

Saffron & Slate has built a strong reputation across Dubai's design community — but in the AI-search era that reputation is invisible. Our GEO audit scored the studio's site 52/100 (Grade C): when prospective clients ask ChatGPT or Perplexity for an interior design studio in Dubai, Saffron & Slate is rarely the answer, while two direct competitors are already being cited.

AI-referred traffic is growing 527% year over year. This proposal lays out a 90-day plan to close that gap and turn AI search into a predictable source of design briefs — with the ROI math, phased scope, and proof to back it. Every month of delay hands the citations to competitors who moved first.

Situation Analysis

Missing. The generic proposal never mentions the client's actual numbers — no audit, no competitors, no evidenced pain. The prospect has no reason to believe the diagnosis.

The site scored 52/100 overall, dragged down by three weak categories from the audit:

CategoryScoreStatus
Schema Markup30/100Critical
Trust Signals45/100Weak
AI Readiness55/100Mixed
Content Quality75/100Strong
Technical SEO85/100Strong

Opportunities: add FAQPage + Organization schema, publish an llms.txt file, restructure top pages into answer-first passages AI can quote directly.

Proposed Solution & Scope

  • SEO optimization
  • Content creation
  • Ongoing support & maintenance

Three vague bullets. No phases, no quantities, no exclusions — an open door for scope creep and mismatched expectations.

WEEKS 1–2Phase 1 — Foundation

Schema fixes (FAQPage + Organization JSON-LD on top 10 pages), llms.txt, crawler access. Milestone: machine-readable to AI systems.

WEEKS 3–8Phase 2 — Growth

4 GEO-optimized pages/mo, answer-first restructuring, monthly re-score report. Milestone: first AI citations.

WEEKS 9–12Phase 3 — Scale

Competitor citation tracking, quarterly strategy review, compounding content plays. Milestone: predictable AI-referred briefs.

  • Exclusions: paid-ads management, full site rebuild, PR / link outreach — quoted separately if needed
  • Client responsibilities: CMS access, brand assets, approvals within 3 business days
  • Cadence: kickoff, biweekly check-ins, monthly review · email SLA 1 business day

Investment

$3,500 per month — marketing retainer.
A naked number: no tiers, no anchor, no ROI context. Take it or leave it.
AuditGrowth ★ RECOMMENDEDDominate
Price$1,500 one-time$3,500/mo$7,000/mo
GEO audit & fixesYesYesYes
Monthly re-scoringNoYesYes
Content productionNo4 pages/mo10 pages/mo
Competitor trackingNoYesYes
Expected 12-mo ROI3x6x9x

Top tier anchors, middle tier is the target, bottom tier is the foot in the door — the engine's mandatory pricing psychology. The gold column is exactly how the PDF renderer highlights ★ RECOMMENDED.

ROI Projection

Missing. No ROI section at all — the price stands naked, and the buyer is left to do the value math alone. Most won't.

Cost of inaction: AI-referred traffic is growing 527% YoY; at 52/100 AI readiness the studio captures almost none of it — an opportunity walking to the two competitors already cited.

ScenarioAnnual value createdAnnual investmentNet ROI
Conservative$126,000$42,0003.0x
Moderate$252,000$42,0006.0x
Aggressive$378,000$42,0009.0x

Engine rule: the conservative case must still be positive — if it isn't, the scope or price is wrong and the agent flags it. All assumptions are stated in the generated document.

Next Steps

Please let us know if you have any questions. We look forward to hearing from you.

No path to yes, no date, no validity — the proposal ends and simply hopes.

  1. Review this proposal and share it with your team
  2. Book a 30-minute walkthrough call — two times proposed in the cover email
  3. Sign via e-signature and we begin within one week

Likely objections ("too expensive", "we have an agency", "does AI search matter?") are pre-handled inside the document — see the objection cards below.

Appendix — follow-up sequence, generated with every proposal:

Day 0 · deliveryDay 2 · walkthroughDay 5 · value-addDay 9 · soft close
The go / no-send gate

Every proposal is scored before it's sent.

This is the engine's actual rubric — five categories at 20% each, every category scored 0–100 across four sub-dimensions by its own agent. Below 70? It goes back for rework, not into the prospect's inbox.

  • 85–100A+Send it — this proposal closes
  • 70–84AStrong — minor polish before sending
  • 55–69BAverage — meaningful gaps to fix
  • 40–54CWeak — likely to lose, rework needed
  • 25–39DPoor — will not win, rebuild
  • 0–24FCritical — start over
Problem-Solution Fit20%
Problem evidence /30Specificity /25Competitive insight /20Buying readiness /25
Pricing & ROI Clarity20%
Tier structure /25Market alignment /25ROI math /30Budget fit /20
Scope & Deliverables20%
Deliverable specificity /30Phasing logic /25Expectation setting /25Scope protection /20
Value & Business Case20%
Value lever clarity /25Cost of inaction /25Three-scenario rigor /30Assumption transparency /20
Persuasion & Close Strategy20%
Executive summary /30Differentiation /20Proof quality /25Objections & close /25
Pre-handled, in writing

The six objections that kill deals — defused before they're raised.

The engine's objection playbook writes the rebuttal into the proposal itself, so the buyer's internal skeptic meets an answer, not a gap. Flip a card to see the pre-handled response.

Not a one-shot generator

Audit to signed SOW — one system.

The proposal is the middle of a closing pipeline, not the end of it. If an audit report is sitting in the folder, the engine auto-detects it and builds on real findings.

Auditauto-detects GEO-AUDIT-REPORT.md & friends as the evidence base
Build/proposal build — 5 agents draft in parallel
Score0–100 composite gate — below 70, rework the weak section
PDF/proposal report-pdf — branded cover, gold ★ tier
Follow-upDay 0/2/5/9 emails, written and ready to send
SOWon acceptance, /proposal sow turns the yes into a contract
What lands in the folder

Markdown for you. A branded PDF for the client.

The cover on the left is a CSS recreation of the actual reportlab theme the generator ships — navy band, sky accent, gold-highlighted ★ RECOMMENDED pricing column inside.

PROPOSAL-[Client].md · scored proposal PROPOSAL-[Client].pdf · client-ready PDF PROPOSAL-FOLLOWUP-[Client].md · 4 emails PROPOSAL-COVER-[Client].md · delivery email PROPOSAL-SOW-[Client].md · on acceptance

Plus pricing-only, scope-only, ROI-only, summary, case-study, objections, compare, upsell and 60-second quick-outline outputs — 14 commands, one orchestrator, 13 sub-skills, 5 agents.

Working with Mina

Cheap insurance against losing one $5k deal.

No packages to pick from — Mina runs the engine on your live prospect, hands you the scored, branded, follow-up-ready proposal, and you send it and take the meeting. Scope and pricing are set to your pipeline on a quick call.

Proposal-to-Close Engine

The next proposal you send
could carry a score.

Pitching something this week? Send the prospect's name and what you're proposing — you get back the scored document, the branded PDF, and the follow-up sequence, usually within 48 hours.

Human-reviewed before anything reaches your prospect. No dashboards to learn, no software to install.