AI lead qualification: screening tire-kickers before they hit your calendar.

It's April in the Metroplex, your phone is finally ringing, and your estimator just spent 90 minutes driving to Rockwall to quote a fence job for someone whose real budget was a third of your minimum. Meanwhile, a genuinely ready buyer in Keller filled out your form, waited two days for a callback, and hired someone else.

That's the spring pattern for most service businesses: demand goes up, but so does noise. And the standard fix — "just book everyone and sort it out on-site" — burns the most expensive resource you have. If your estimator's loaded cost is $50/hour and a wasted quote trip eats two hours with drive time, every dead appointment costs you around $100 before you count the real job you didn't get to.

The fix isn't working harder. It's putting a qualification layer between your contact form and your calendar.

Why "book everyone" fails in spring

Your calendar is a fixed asset. In March through June, it's your scarcest asset. Every slot a tire-kicker takes is a slot a real buyer can't have — and speed matters more than most owners think. A widely cited Harvard Business Review study (Oldroyd, McElheran, and Elkington, 2011) found that companies contacting leads within an hour were roughly seven times more likely to qualify them than those waiting even an hour longer. So the game isn't just filtering bad leads; it's fast-tracking good ones.

The old-school gatekeeper was a sharp office manager who asked three questions before booking anything. Most small operations don't have that person anymore, or she's slammed. The qualification layer is that person, automated, working at 9pm on a Sunday when half your leads actually come in.

The three questions that separate real jobs from noise

You don't need twenty fields. You need three answers before anyone sees your calendar:

1. Budget. Not "what's your budget?" (people dodge that) — offer ranges. "Most projects like this run $X–$Y. Does that fit what you had in mind?" An illustrative example: a typical Keller plumber doing repipes might set the floor at $3,000. Anyone who picks "under $1,000" gets a polite email with a referral or a DIY resource — not a truck roll.
2. Timeline. "Ready to start in the next 30 days," "1–3 months," or "just researching." Researchers aren't garbage — they're future customers. But they get a follow-up email sequence, not a Tuesday-morning estimate slot.
3. Address. The quiet killer. A Fort Worth crew quoting in Forney spends more on windshield time than the margin on a small job. Set a service radius and enforce it at intake. Outside the radius? Either an automatic "here's our trip fee" disclosure or a courteous decline.

If a lead clears all three, then they see the calendar — ideally instantly, while intent is hot.

Two ways to build it: conditional forms vs. conversational AI

Option 1: Conditional (branching) forms. Tools like Typeform, Jotform, or a custom-built form show questions based on prior answers, score the lead, and only route qualified leads to your scheduler (Calendly, Housecall Pro, ServiceTitan, whatever you run). This is the cheap, reliable version. You can ship it in a week.

Weakness: forms feel like forms. Rigid branching can misclassify edge cases — the commercial client whose "budget" question doesn't map to your residential ranges, for example.

Option 2: Conversational AI intake. Instead of a form, the lead chats — on your site, or by text after they submit a short form. The AI asks the same three questions, but naturally, and it can handle the messy middle: "Well, insurance might cover part of it" or "I'm at my rental property in Arlington but I live in Weatherford." It extracts budget, timeline, and address from ordinary sentences, scores the lead, and either offers a booking link on the spot or routes to a human.

The conversational version costs more to build and needs guardrails (it must never quote prices, promise scope, or book outside your rules), but it converts better because it feels like talking to that sharp office manager — and it answers at midnight.

A pragmatic middle path: short form up front (name, service, zip), AI-driven follow-up by SMS for the qualifying questions. Texts get read; long forms get abandoned.

Calibration: filter tire-kickers without losing real buyers

This is where most qualification setups die. Set the bar too high and you're screening out good customers who just answered a question awkwardly. Rules that keep it honest:

• Never dead-end anyone. Unqualified doesn't mean unworthy. "Under budget" leads get a helpful email; "just researching" leads get a monthly tips sequence. Some of them come back ready.
• Keep it to 3–5 questions max. Every additional field costs completions. You're qualifying, not underwriting.
• Let borderline cases through — flagged. If someone's answers are ambiguous, book them, but mark the appointment so your estimator can do a two-minute phone screen first instead of a drive.
• Review misses monthly. Pull the last 30 days: which booked appointments turned out to be duds, and did their intake answers predict it? Which declined leads called back angry or hired a competitor for a real job? Tighten or loosen one rule at a time.
• Track two numbers: percentage of appointments that produce a legitimate quote opportunity, and form-to-booking completion rate. If the first goes up and the second doesn't crater, your calibration is working.

Expect to tune it for a month or two. That's normal, not failure.

How West Fork approaches this

We build qualification layers both ways — branching forms wired into your existing scheduler for a fast, low-cost win, or a conversational AI intake (web chat or SMS) when the volume justifies it. Either way, we start with your real numbers: your minimum job size, your service radius, your close rate by lead source. Then we calibrate against a month of actual leads before locking rules in. No black boxes — you can see and change every rule.