An 8-step proposal walkthrough on a real (fake) engagement

Most articles about AI for proposal writing stop at the framing. They tell you AI can help. They tell you to write a brief first. They tell you to edit the output. They don't show you the actual chain of prompts running on actual material, end to end, producing the actual artifact you'd send to the actual prospect.

This article does that. The setup is fictional but plausible — a fictional B2B SaaS company, a fictional discovery call, a fictional consultant. Everything else — the prompts, the inputs, the outputs at each step, the way each step's output feeds the next — is the real workflow we run, taken out of Chapter 3 of The Solo Consultant's AI Playbook and the Proposal-Closer Prompt Pack.

If you've read the previous articles on proposals and why prompts feel generic, this is the worked example that shows the principles running. The frame is in those articles; the run-through is here.

The setup

You're a solo consultant who helps B2B SaaS companies fix broken onboarding funnels. You just had a 30-minute discovery call with Priya Shah, Head of Product at Acme Widgets — a $4M ARR SaaS doing bookkeeping automation for small restaurants. Priya said their trial-to-paid conversion dropped from 18% to 9% over the last quarter and they don't know why.

Total time, end to end with the prompt chain: about 25 minutes. Without it: 90+ minutes for a proposal that's still slightly worse.

Step 1 — Discovery debrief (3 minutes)

Right after the call, paste your raw notes into the debrief prompt. Notes can be messy bullets — that's the point. The model produces a structured one-page debrief in 30 seconds.

Inputs you give it:

• Priya, Head of Product, design background
• Acme Widgets: bookkeeping auto for small restaurants, ~$4M ARR
• Trial conversion was 18%, now 9%
• Changed onboarding copy 2 months ago, added video walkthrough
• Engineering lead thinks top-of-funnel quality issue, Priya disagrees
• Nobody has looked at session recordings since January
• They have Mixpanel but "nobody really uses it"
• Budget: "we have some budget, just want someone who knows what they're doing"
• CEO wants conversion back to 15%+ by end of Q3 (~10 weeks away)
• Priya alone on decision, CEO signs off

The debrief output identifies what Priya actually needs (a diagnostic, not a copy rewrite — she's looking for an outside read to resolve the internal disagreement with engineering), the budget signal (soft positive, no number), the decision process (Priya decides, CEO signs off, ~2-week timeline), three red flags (internal disagreement risk, weak data hygiene, possibly unrealistic CEO expectation), and a recommended next step (send proposal, but ask one budget-range question first).

That's 30 seconds of generation that gives you a debrief sharper than what you'd write by hand at 4 PM Friday.

Step 2 — ICP fit check (1 minute)

You've already saved your ICP criteria once (Proposal-Closer Pack Prompt 02 has a template). Paste those criteria + the debrief into the fit-flag prompt. Output: GREEN, 80% confidence, three reasons.

The verdict isn't "this will close." It's "this is worth proposing for." Different thing, and the distinction is what saves you the four hours per quarter you'd otherwise spend writing proposals for deals that were never going to close.

For Acme Widgets, GREEN. Onward.

Step 3 — Scope generator (3 minutes)

Inputs: the debrief, the ICP-pass result, your hourly floor ($175), available hours per week (15). The scope generator outputs three differentiated options:

• Option 1 — Conversion Audit (Small). 2-week diagnostic. 18 hours. Documented current funnel, ranked top-3 fixes. For the prospect who wants clarity fast, not transformation.
• Option 2 — Conversion Turnaround (Right-Sized) ← Recommended. 6-week engagement. 55 hours. Diagnose, fix, train. Includes session-recording review, copy rewrites, weekly check-ins, end-of-engagement playbook.
• Option 3 — Conversion + Activation System (Expansive). 10-week engagement. 85 hours. Above plus a 30-day activation-metrics framework, two team workshops, post-engagement async support.

The differentiation matters. Three undifferentiated scopes ("4 weeks / 6 weeks / 8 weeks") force the prospect into decision-load you're charging them for. Three meaningfully different scopes give them a clear recommendation with two real alternatives.

Step 4 — Pricing anchor (2 minutes)

Inputs: the three scopes, the budget signal from the debrief ("we have some budget"), your hourly floor. The pricing anchor produces three prices with the anchor logic attached:

• Option 1: $5,500. Anchored against the cost of a senior product-analytics hire ($180k/yr fully loaded). Two weeks of senior output for ~3% of that annual cost.
• Option 2 (recommended): $14,700. Anchored against the recovered ARR. Lattice loses ~$360k/yr at 9% conversion vs. the 18% baseline. Engagement pays back in three weeks of recovered conversion at the 15% target.
• Option 3: $26,500. Anchored against avoiding the next consultant hire when conversion dips again.

The anchor logic is the whole point. A price without an anchor story is a number the prospect has to defend to their CFO with no help from you. A price with a one-sentence anchor is a number they can copy into the email forwarding the proposal.

Step 5 — Proposal draft (5 minutes)

Inputs: everything above, plus client details and your name. The draft output is a complete proposal in the structure from the proposal-shape article — eight sections, under 500 words.

The opening paragraph reads like the second paragraph of a memo from a peer who's been sitting with the problem since the call:

Priya — you've watched trial-to-paid drop from 18% to 9% over the last quarter, and there's an internal split on why. Your engineering lead thinks top-of-funnel; your gut says something changed further down. Both could be true; more likely one is dominant and the other is a distraction. You don't need more guesses — you need a diagnostic you can act on before Q3 ends.

That's a paragraph the prospect's CEO will read twice. It signals you understood the political dynamic, you've taken a position, and you have a specific path forward — all without saying "thank you for the opportunity" once.

You'd edit one or two phrases to match your voice precisely (the "you don't need more guesses" cadence is Claude's, not yours), but the structural work is done. The 90-minute version of this proposal is now 5 minutes plus a 5-minute edit pass.

Step 6 — Objection handler (when Priya replies, 2 minutes)

Two days later, Priya replies. Her CEO is pushing back on the $14,700 — "he's comparing it to the $8k we paid a freelancer last year who didn't deliver." You have 12 minutes between meetings to write a response that doesn't sound defensive.

The objection handler outputs three options:

• Stand firm. Reframe what's different about this engagement vs. the cheap freelancer ("the difference isn't hours, it's what you're left with").
• Reasonable flex. Drop $2k in exchange for the right to use the engagement (anonymized) as a case study.
• Walk away gracefully. Polite exit that keeps the door open for later.

Each option includes the email response, what you're signaling by choosing it, and when to pick it.

For Acme Widgets, the flex option fits — Priya needs ammunition for the CEO conversation, the $2k cut buys you a future case study, and the discount-for-rights trade reads as professional rather than weak. You send Option B. Three minutes total including the choice.

Step 7 — Follow-up sequence (queue it now, 2 minutes)

You hear nothing for a week. You don't want to nag, but you don't want to let the deal die. The follow-up prompt outputs three emails for day 3, day 7, day 14 — each escalating in directness without escalating in desperation.

Day 3: a value-add nudge with a related resource. Day 7: the direct "is this a fit or should I close the file?" email. Day 14: the take-away close — "closing the file on this one" — counterintuitively the highest-reply email of the three.

You queue them. They send themselves on schedule. The deal either re-engages or doesn't, and you're not the one chasing.

For Acme Widgets, Priya replies to the day-3 nudge: CEO approved, kickoff next Monday, final price $12,700 with case-study rights as agreed. Deal closed.

Step 8 — Deal debrief (after wrap, 3 minutes)

The single most important prompt in the chain, and the one most consultants skip. After the engagement wraps — won, lost, or stalled — you run the debrief on the whole deal cycle.

Outputs include: what specifically won (or killed) the deal, did you price it right, did you scope it right, ICP signal updates, prompt upgrades for next time, and the one line to remember.

For Acme Widgets, the one line might be: "When a client compares you to a cheap vendor, don't defend the price — reframe the category."

Without this prompt, every deal teaches you nothing. With it, every deal teaches you something specific you can apply on the next one. The compounding is real and invisible — the deals you close in year three look different from year one because of all the debriefs in between.

Total time

End-to-end on a real proposal: ~25 minutes including edit passes. Without the chain: 90+ minutes for a slightly worse output, and the deal-debrief part doesn't happen at all because there's no prompt forcing the discipline.

The chain produces proposals that close at higher rates than the old way of writing them, and produces them faster. Both axes matter. Most AI-for-proposals articles make you choose between the two; this one is the structural argument that you don't have to.

Going deeper

Every prompt in this walkthrough is in the Proposal-Closer Prompt Pack at $37, with the full Acme Widgets worked example, a drop-in Claude skill, and ChatGPT-compatible variants on every prompt. If proposals are a recurring bottleneck, the pack is the deeper version of this article.

The Pack pairs with The Solo Consultant's AI Playbook for the broader operating model. The bundle is at digitalkreative.co/products/solo-consultant-starter-bundle at $67, $17 off standalone.

— Digital Kreative