How I Used Claude to Close a $20K Project

I almost lost a €20,000 consulting project because I sent a proposal that read like a form letter. The client — a developer looking to sell six luxury apartments in Funchal — told me directly: “This doesn’t feel like you understand our situation.” He was right. I’d copied a structure I’d used a dozen times and swapped in the address. That was January 2026, and it was the kick I needed to finally build a proper proposal process around Claude.

What followed over the next three weeks changed how I run every high-value deal. I closed that project. I also learned exactly where Claude earns its place in a solo real estate operation — and where it falls flat.

The Problem: Solo Operators Don’t Have Time for Custom Everything

Running a one-person real estate consultancy in Madeira means I wear every hat. I’m the agent, the marketer, the proposal writer, the follow-up machine. For standard listings — a two-bedroom apartment in Câmara de Lobos, a countryside quinta — I have templates that work fine. But the moment a project gets complex, those templates betray me.

The Funchal luxury development was exactly that kind of deal. Six units, mixed typologies, a developer client who’d worked with agencies before and was skeptical of solo consultants. He wanted evidence I understood the buyer profile, the pricing dynamics of Funchal’s upper market, and how I’d position six units without cannibalizing each other. That’s not a template problem. That’s a thinking problem. And thinking takes time I don’t always have.

I’d been using Claude (Anthropic’s AI, currently on Claude 3.7 Sonnet via the Pro plan at $20/month) for property descriptions and email drafts since mid-2025. But I’d never used it systematically for a deal this size. This project pushed me to build a real process.

What I Actually Tried: Building a Claude-Assisted Proposal System

My goal wasn’t to have Claude write the proposal. My goal was to use Claude to think faster and more clearly than I could alone — then write the proposal myself with that foundation under me.

Step 1: The Client Intelligence Brief

Before I wrote a single word of the proposal, I fed Claude everything I knew about the client. Notes from our two meetings, the developer’s previous project (a mid-range block in São Martinho I found via the Conservatória do Registo Predial records), the pricing on comparable Funchal luxury units I’d pulled manually, and three questions he’d asked me that I hadn’t answered well in the room.

I asked Claude: “Based on these notes, what does this client most likely fear about hiring a solo consultant over an agency? What objections haven’t been raised but probably exist?”

The output was sharp. Claude identified three unspoken concerns: capacity (could I handle six units simultaneously?), network (did I have access to the buyer segment he was targeting?), and accountability (what happens if sales stall?). I hadn’t framed any of those explicitly, but they were all sitting in my notes. Claude surfaced them faster than I would have on my own.

That took about 25 minutes, including two rounds of follow-up prompts.

Step 2: The Positioning Argument

Once I had the objections mapped, I used Claude to stress-test my own positioning. I typed out my intended argument — why a solo consultant is actually better than an agency for a boutique luxury development in Funchal — and asked Claude to argue against it. Play the skeptical developer. Find the holes.

It found two real weaknesses. My claim about “personal attention” was vague and every agency says the same thing. My reference to “deep local market knowledge” needed evidence, not just assertion. Claude suggested I anchor both claims in specifics: names of comparable sales I’d personally handled, response time commitments with numbers attached, a named buyer database rather than a generic “network.”

Good feedback. The kind a business partner would give if I had one.

Step 3: The Proposal Structure and Section Drafts

I asked Claude to suggest a proposal structure specifically for a developer selling luxury residential units in a European island market, where the buyer pool is international and the competition for listings comes from multi-agent firms. It generated a solid 8-section outline in about 90 seconds. I cut it to 6 sections. Then I drafted each section myself using Claude’s competitor objection analysis as my guide.

For the two sections I was least confident about — the market analysis summary and the fee justification — I asked Claude to write first drafts based on the data I provided. I rewrote both almost entirely, but having a draft to react to is twice as fast as writing from nothing. That’s a consistent pattern I’ve noticed across dozens of tasks: rewriting is faster than originating.

Step 4: The Follow-Up Sequence

After sending the proposal, I used Claude to build a 3-email follow-up sequence. Not generic check-ins. Each email was written around a specific value-add: a market data point, a question that would open conversation, a low-pressure nudge with a clear next step. I gave Claude the proposal content and the client profile, asked for three emails spaced 4, 9, and 16 days out, each under 120 words.

The client responded to the first follow-up. We signed the consulting agreement 11 days after I sent the proposal.

My Real-World Experience: What the Numbers Actually Look Like

Let me be specific about the time this took, because I tracked it.

The previous time I’d built a proposal for a comparable-complexity project — a 4-unit boutique development in Calheta, back in late 2024 — I spent roughly 11 hours across four days. Research, writing, revising, second-guessing the positioning, rewriting the intro three times. I closed that deal too, but the proposal process was genuinely painful.

For the Funchal project with my Claude-assisted workflow, I tracked every session. Here’s the breakdown:

Task	Time Without Claude (estimated)	Time With Claude (actual)
Client intelligence brief	2.5 hrs	25 min
Positioning stress-test	1.5 hrs	30 min
Proposal structure + section drafts	5 hrs	2.5 hrs
Follow-up email sequence	1.5 hrs	20 min
Final review and polish	1 hr	45 min
Total	~11.5 hrs	~4.5 hrs

Seven hours saved on a single proposal. At my consulting rate, that’s meaningful. More importantly, the output was better — tighter argument, sharper positioning, specific rather than generic. The client said so directly in our signing call: “Your proposal addressed things I hadn’t even asked you about yet.”

That’s the part I couldn’t have manufactured without the client intelligence brief Claude helped me build. The objection mapping changed the proposal’s logic from the ground up.

Since that project, I’ve used this same four-step process on two other consulting engagements — one in March 2026, one in April. Both closed. I’m not claiming Claude closed the deals. I closed them. But the thinking infrastructure Claude helped me build meant I walked into those conversations with better preparation than I’ve had in years.

Where Claude Genuinely Struggled in This Process

Here’s the honest part.

Claude has no idea what’s actually happening in the Funchal luxury market right now. It doesn’t know that a specific new development in the Lido area is affecting buyer expectations, or what prices per square meter moved to in Q1 2026. Its training data has a cutoff, and even with web browsing enabled in some configurations, I found the market-specific data unreliable. I couldn’t use any Claude-generated market figures without verifying them against sources I trust: INE Portugal, Confidencial Imobiliário, my own transaction records.

In one session, Claude confidently cited a price trend that was directionally correct but about 18 months out of date. If I’d included that in the proposal without checking, I’d have looked sloppy in front of a developer who knows the market cold. I caught it because I know the numbers. A less experienced consultant might not have.

The other limitation: Claude can’t read a room. The nuance of what my client actually cared about — the things he said sideways, the hesitation when I mentioned the fee structure — came from two in-person meetings. Claude worked with my interpretation of those signals, not the signals themselves. The quality of its analysis was entirely dependent on the quality of my notes. Garbage in, garbage out still applies.

How I’d Set This Up If You’re Starting From Zero

If you’re a solopreneur trying to apply this to your own high-value proposals, here’s what I’d actually recommend:

• Get Claude Pro ($20/month). The free tier works for simple tasks, but for long proposal sessions with multiple files and extended context, you’ll hit limits at the worst moment. Pro gives you enough headroom to work through an entire proposal process without interruption.
• Start with objection mapping, not writing. The biggest mistake is going straight to “write me a proposal.” Make Claude help you think first. Feed it everything you know about the client and ask what they’re afraid of. That changes everything downstream.
• Use it to argue against you. Ask Claude to steelman your competitor’s position or attack your own proposal. This is where it earns its cost on complex deals. You need someone to push back, and solo operators rarely have that.
• Verify every number independently. Especially any market data. Claude’s knowledge has limits and it doesn’t always signal when it’s out of date.
• Write the final version yourself. Your voice, your relationships, your credibility are what close high-value deals. Use Claude to build the scaffold. You put up the walls.

My Rating and Why

Claude for high-value proposal work: 4.2/5. It earns that because the objection-mapping and positioning stress-test features directly improved a proposal that closed a €20,000 project — but the market data limitations and its inability to replace real client relationship context keep it from a higher score.

Practical Summary: What to Take From This

I saved seven hours on one proposal. That proposal closed a €20,000 project. The process cost me $20 that month in Claude Pro fees plus about 4.5 hours of focused work.

The ROI math is obvious. But the less obvious win is that the quality went up, not just the speed. Faster and better is rare. With the right workflow, Claude delivers both on proposal work.

This only works if you put in good inputs. Your meeting notes, your market knowledge, your read on the client — Claude amplifies those. It doesn’t replace them. If you walk in with shallow preparation, Claude will produce a polished version of shallow thinking.

Walk in prepared, and it gives you a thinking partner that’s available at 11pm when you’re staring at a blank page with a proposal due in two days.

If you want to see the exact prompts I used for the client intelligence brief and the objection-mapping session, I’ve put them together in a short guide — grab it here. These are the actual prompts from the Funchal project, not templates built for an imaginary client.