The highest-quality inbound leads for home service businesses in 2026 are being generated by AI search — ChatGPT, Perplexity, and Google's AI Overviews. Invoca's 2026 analysis of 70 million home service calls found that AI search referrals post a 45% lead rate, the highest of any marketing channel measured. But 48% of those calls go unanswered (Invoca, 2026).
Why AI search callers arrive differently
When someone searches Google for "HVAC repair Spokane," they see results and pick one. When someone asks ChatGPT for a heating contractor recommendation, they've had a back-and-forth first — described the problem, asked follow-up questions, compared options — before they ever touch the phone. By the time they dial, they're not comparison shopping. They've already made a tentative decision.
That pre-vetting shows up in the data. A "lead" in Invoca's framework is a call where the caller expresses clear buying intent: needs a quote, wants to schedule, is ready to commit. Across all home service channels, 38% of answered calls qualify as leads. For AI search referrals, that number is 45% — five points above paid search (Google Ads), which posts 40%. AI search callers are self-filtering before they call (Invoca, 2026).
Volume matters here. AI search is still a small share of total inbound call volume compared to Google Ads or organic search. But the quality advantage is real and measurable right now, and that share is growing as more consumers fold ChatGPT and similar tools into how they find local services.
The 48% answer gap
Invoca's 2026 home services analysis found an average 52% call answer rate across nine sub-industries — meaning about 48% of all inbound home service calls never reach a person (Invoca, 2026). For a Spokane HVAC company, plumber, or general contractor running any kind of digital marketing, that math is unfavorable regardless of which channel is generating the call.
If your business generates 80 inbound calls per month and 48% go unanswered, you're paying to send roughly 38 callers to voicemail. Some will call back. Most won't — they'll move to the next result on their list. The AI search callers with the 45% lead rate hit the same gap as everyone else. That means the highest-intent segment of your inbound volume is also the one most likely to be permanently lost when it goes unanswered.
The channel comparison: where lead quality actually comes from
| Channel | Lead Rate | Notes |
|---|---|---|
| AI Search (ChatGPT, Perplexity, AI Overviews) | 45% | Highest quality; caller pre-vetted through AI conversation before calling |
| Paid Search (Google Ads) | 40% | High volume, strong intent; currently the largest share of inbound calls |
| All channels (home services average) | 38% | Cross-channel baseline from 70M+ calls analyzed |
Source: Invoca, 2026 — 70 million home service calls across 9 sub-industries and 7 marketing channels.
The second gap: answered but not closed
The answer rate problem has a second layer. Of the calls that do reach a person, many don't result in a booked appointment — not because the caller wasn't interested, but because no one asked. Invoca found that 55% of home services businesses don't actively request a booking commitment during the call (Invoca, 2026). The caller expressed interest, talked to someone, and left without a scheduled job.
The full funnel for a Spokane contractor, assuming 100 inbound calls:
- 52 reach a person — 48 go unanswered
- 38% of answered calls are leads: roughly 20 qualified callers
- 45% of those leads convert during the call: about 9 booked jobs
- 91 calls didn't become a booked job — including 48 the business never spoke to
A five-point improvement in each of those three rates — answer rate, lead rate, on-call conversion — yields roughly 38% more revenue from the same inbound call volume, according to Invoca's modeling (Invoca, 2026). The phone doesn't need to be perfect to move the number. It needs to be better at each step.
What changes when AI answers those calls
The answer rate gap closes when the phone is always answered. An AI system handles inbound calls 24/7 — standard inquiries, booking requests, after-hours calls. For Spokane contractors, that's particularly relevant in winter: a furnace failure at 10pm in January is a high-stakes call that currently hits voicemail at most shops. The caller who reaches a live response gets their appointment booked. The one who hits voicemail calls the next name on the list.
The booking gap closes differently. When a booking request is built into every AI-handled call flow — not left to whether a staff member thinks to ask — the 55% who don't get a close becomes a configuration decision rather than a training problem. The AI gathers availability, checks the schedule, and confirms the appointment before the call ends. Complex calls route to a human immediately.
The less obvious change is the data layer. Every call captured by an AI system is logged, transcribed, and categorized: call time, duration, caller questions, intent signals, booking outcome. For a Spokane contractor running Google Ads and also appearing in ChatGPT recommendations, that means you can see which calls came from which channels and which ones actually converted. Marketing attribution that most home service businesses don't have — because voicemail doesn't tell you anything.
AI search is generating your best leads now. The question is whether AI on the receiving end is there to catch them.