Local services with after-hours leads: AI books the appointment, you find the booked job in the morning

The after-hours lead is the easiest revenue to lose

Local-service businesses live and die on lead capture. Public data from HomeAdvisor, ServiceTitan, and Yelp lines up on a hard truth:

• 78% of leads go to the first business that responds. On after-hours, that's rarely you.
• Average response time for a missed call in residential services is 12+ hours. The customer has already booked elsewhere.
• 30-50% of after-hours inquiries are routine bookings, not emergencies. The customer just happens to be free to think about their broken dishwasher at 8pm.

OpenAgent isn't replacing the after-hours technician. It's replacing the answering machine.

How the AI handles a typical request

Configuration is light. A knowledge base of your services, your pricing ranges, your service area, and your hours; an escalation rule on emergency keywords; a workflow that captures booking requests with contact details. Total setup: 30-45 minutes.

A walkthrough of a routine HVAC call:

• 9:42pm Saturday. Visitor lands on your site, finds the chat bubble. “My AC stopped working today.”
• AI: “Sorry to hear that. Quick question: is it just not cooling, or is it not turning on at all? And what's your zip code so I can check if you're in our service area?”
• Visitor: “Not turning on. 30033.”
• AI: “Got it, you're in our area. We can send a tech out Sunday morning or Monday afternoon. Service call is $90, waived if you book a repair. Want me to pencil you in?”
• Visitor: “Sunday morning please.”
• AI: “Done. Can I grab your name, phone, and address?”

Booking lands in the dashboard with the conversation transcript attached. You see it Sunday morning and confirm with the customer.

Emergency escalation: the right friction

Half of your business value as a local-service operator is being there for emergencies. The escalation policy uses keyword + intent classification so the right calls page you and the wrong ones don't:

• Pages you immediately: “water everywhere”, “no heat”, “gas smell”, “tooth pain”, “severe”, “emergency”, “can't breathe”.
• Schedules for morning follow-up: routine bookings, price questions, scheduling preferences, anything that isn't time-critical.

The page lands via Slack, browser push, SMS (with Twilio webhook), or all three. Median latency from intent classification to page: under 5 seconds.

Sample math: a small HVAC business

Defaults below are sized for a single trade location. Multi-location operators scale every output proportionally; the model handles the increase without any extra cost on the OpenAgent side.

Current monthly after-hours bookings: 800 × 30% × 3% = ~7 bookings.

With AI chat: 800 × 30% × 10% = ~24 bookings.

Incremental bookings/month: ~17. Incremental revenue: 17 × $400 = $6,800/month. Incremental profit: $6,800 × 35% = $2,380/month.

Annual incremental profit: ~$28k on a single trade location. Multi-location operators (a dental group, a plumbing chain) see this scale per location.

Lead quality, not just quantity

A common worry: “more leads via chat means more junk to triage.” Two design choices make the lead quality higher than form-fill:

• Qualification before booking. The AI asks about the problem, the location, and the urgency before it offers a slot. By the time the booking lands you already know it's real.
• Spam filtering. The same intent classifier that triggers escalation also flags obvious spam (offers to redesign your website, SEO pitches, etc.) and never escalates them. They land in a separate folder.

Why this matters more for local services than for ecommerce

Ecommerce's ROI from chat is real but secondary; the customer could have bought anyway. Local services is different: the customer is calling because something is broken right now, and the first responsive business wins the job. A 7-percentage-point conversion lift on after-hours sessions is a bigger swing than almost any other marketing or operational change available to a small-trades operator.