The Math You're Ignoring
You're losing $24,000 to $32,000 annually on discovery calls with tire kickers. Research shows 50% of sales time vanishes on unqualified prospects, and for solo consultants specifically, 30-40% of business development cycles burn on leads that won't convert. That's inventory—time—you'll never recover. The fix isn't saying no to more people. It's letting your intake system say no for you.
The Dan Kennedy Rule That Changed My Fees
Dan Kennedy taught me something I never unlearned: the intake IS the sale. If you're explaining your rates on a call, your intake failed. I watched my first $50K engagement die in week three because I'd never asked the prospect about budget during intake. Zero qualification questions. Just assumed. The discovery call was theater,they were already disqualified. A self-qualifying intake system fixes this by asking the hard questions before you ever dial. No surprises on the call. No post-close negotiations. The deal structure is locked.
FOCUS Strategy: Your Intake Doctrine
Use this framework to design your system:
F , Filter ruthlessly. Disqualify early and often. Bad fit is worse than no fit. Budget check, authority check, timeline check,three questions. If any lands soft, they're not ready.
O , Organize the data. Intake form captures: budget, authority, implementation timeline, current pain (quantified), decision-making process, and competitive context. Unstructured answers become useless. Every field must map to a qualification threshold.
C , Calculate scope automatically. Your LLM reads their responses and estimates hours. Small scope goes to fast-track pricing. Medium scope triggers standard SOW. Large scope flags you for a prep call before the main discovery.
U , Update the SOW in real-time. As they answer intake questions, the system populates a draft SOW with deliverables, timeline, and assumptions baked in from their answers. They see it 24 hours before the discovery call.
S , Score each lead on a 1-10 scale. Traffic light ranking: red (not ready), yellow (qualified but early-stage), green (ready to close in one call). You only take meetings with yellow and green leads.
The Stack: Weekend-Ready
Option 1: HighLevel (GHL) + Claude API
GHL handles the intake form, lead routing, and CRM storage. You embed a form on your site with conditional logic,if they say budget under $5K, they skip to fast-track track. If they say "not sure yet," the system tags them for nurture and doesn't schedule a call.
Connect GHL to Claude API via n8n or Zapier. When the form submits, it fires the data into Claude. Claude scores the lead (1-10), estimates hours, and generates a draft SOW. The SOW populates a Google Doc template. GHL sends it to them immediately with a calendar link for their discovery call,but only if they scored 6+. Lower scores get a nurture email sequence instead.
GHL automation for professional services supports conditional workflows, so this runs entirely hands-off.
Option 2: n8n + OpenAI/Claude + Zapier
n8n is the operations center. It listens for new submissions from your form (Typeform, Jotform, Gravity Forms,doesn't matter). When a prospect submits, n8n wakes up.
First stop: data extraction and validation. n8n cleans the data and sends it to Claude with a system prompt that includes your qualification rubric. Claude reads the answers and returns: lead score (1-10), estimated project scope (small/medium/large), key risks, and a one-sentence fit assessment.
Second stop: SOW generation. If score is 6+, n8n triggers a template in Google Docs (or Airtable) and populates it with the engagement scope, deliverables, timeline, and rate structure. The system auto-calculates fees based on scope.
Third stop: calendar and email. Zapier books them into your calendar for discovery (if score 6+) or into a follow-up sequence (if score 3-5) or closes the loop (if score under 3) with a "not the right fit" email that's so good they refer someone.
n8n's AI workflow integration enables this end-to-end without custom code.
Option 3: Zapier + Make + LLM
Zapier catches the form submission. Make handles the conditional logic and API calls. OpenAI scores and generates. This is lighter-weight but less flexible,Zapier's pricing compounds if you have many steps, and Make's LLM integration requires more manual prompt engineering.
Go this route only if GHL and n8n feel like overkill.
The Build: 8 Hours Start to Finish
Hour 1: Design your intake form
Question set:
- What specific problem are you solving? (open-ended, feeds the SOW)
- What's your budget range? (hard number or range)
- Who makes the final decision? (you or a committee?)
- When do you need to start? (month/quarter)
- What's your current vendor situation? (context for scope)
- How did you find us? (attribution)
Five questions. Three minutes to answer. No fluff.
Hour 2: Build the GHL/n8n intake flow
If using GHL: Set up the form in GHL. Create a workflow that triggers on form submission. Add conditional logic: route high-fit leads to "ready" and low-fit leads to "nurture."
If using n8n: Create a new workflow. Set webhook as trigger. Add HTTP node to receive form data. Map the fields. Set up the LLM node (Claude or OpenAI) that will score the lead.
Hours 3-4: Design your scoring rubric and prompt
Your system prompt to Claude should look like this:
You are a client qualification expert for [Your Consulting Firm].
Score this prospect 1-10 based on these criteria:
- Budget alignment (must be $X min): /5
- Decision authority (can they say yes?): /3
- Timeline urgency (need to start within 90 days?): /2
Respond in JSON:
{
"score": [1-10],
"budget_fit": true/false,
"authority_fit": true/false,
"timeline_fit": true/false,
"estimated_hours": [number],
"scope_category": "small/medium/large",
"key_risks": "[one sentence]",
"recommendation": "schedule/nurture/pass"
}
Test it with five fake submissions. Tweak the thresholds. Get comfortable with how Claude interprets ambiguous answers.
Hours 5-6: SOW template automation
Build a Google Doc template with placeholders: {{prospect_name}}, {{company}}, {{problem_statement}}, {{deliverables}}, {{timeline}}, {{rate}}, {{total_fee}}.
Create a second n8n workflow (or Zapier path) that fires if the lead scores 6+. It pulls the prospect's answers and Claude's analysis, maps them to the SOW template, and generates a new Doc. The link gets sent to the prospect via email.
Hours 7-8: Calendar routing and email sequences
Calendar: If score 6+, book discovery via Calendly. If score 3-5, send a email explaining timing and ask them to apply in 90 days. If score under 3, send a thoughtful rejection email ("we're not the right match, but here's what you should look for in a consultant") and a referral request.
Email sequences: Use Zapier or GHL's native email builder. Personalize with their company name and stated problem. The tone matters,remember, you're filtering, not convincing.
FAQ
Q: Won't an AI-generated SOW look cookie-cutter?
A: It's a draft. You refine it before the call. But 80% of the scaffolding is done. You're not writing from scratch; you're editing. That's a five-minute task instead of thirty.
Q: What if the LLM scores wrong?
A: Review the first ten prospects by hand. Adjust your prompt. LLMs get better with examples. After twenty, you'll trust it. Keep a human override button,if your gut says yes and the AI says 4, you can bump them to discovery manually.
Q: Do I need to change my rates or SOW structure?
A: No. Your intake system doesn't force a pricing change. It just enforces your existing criteria. If you price by scope, the system estimates scope and calculates fees. If you use flat rates, it applies the flat rate. The system matches your business model.
Q: What if prospects push back on the intake questions?
A: Good. That's data. Someone who won't answer "What's your budget?" isn't a real prospect. The friction filters them. Your discovery calls get 20% shorter because the hard questions are already asked.
Q: Can I deploy this with no code?
A: Yes. GHL and Zapier require zero code. n8n is visual too,you build workflows, not algorithms. But you do need to understand how to map data between apps and write a clear system prompt for the LLM. That's technical thinking, not coding.
Q: How much does this cost?
A: GHL is $99-$300/month depending on tier. n8n's self-hosted version is free; cloud is $25+/month. Zapier is $20-$300/month based on tasks. Claude API is $0.003 per 1K input tokens. For 100 intake forms a month, you'll spend $8-25 on LLM calls. Total monthly cost: $130-400. Break-even is one qualified deal per quarter.
Doctrine Connection
Responsibility beats excuses. Every prospect who wastes your time is a choice you made in intake design. Stop outsourcing qualification to chance and conversation. Build the system. It's your job to disqualify fast and fairly. The system does that better than you can.
Go Build
You have everything you need. GHL's professional services automation exists. n8n's AI integrations are production-ready. Claude's API is stupid cheap. Your intake form is in one of three tools,form it today, integrate it by tomorrow, score your first prospect by Sunday night.
One weekend. No consulting, no waiting. You'll never take another unqualified discovery call again.
,-
Further Reading
How to stop wasting time on unqualified leads
The hidden cost of chasing unqualified leads
AI client intake automation for professional services
Dapta AI intake software guide
Perspective AI's ultimate guide to AI intake
*Jeff Barnes, MBA has no personal position in any company, tool, or platform named in this article. DEMG.ai has no current commercial relationship with any party mentioned. DEMG provides marketing education and systems, not investment advice. Past performance does not guarantee future results.*