In April 2026, a New York startup called Avoca AI raised more than $125 million across its seed, Series A, and Series B rounds, hitting a $1 billion valuation. Kleiner Perkins anchored the early rounds; Meritech and General Catalyst led the latest one. The product is not a foundation model, a coding agent, or anything you'd put on a keynote slide. It answers the phone for plumbers, HVAC contractors, roofers, and electricians when nobody else can get to it.
I build AI automation systems for small businesses for a living, and I've written more than once on this blog about missed-call text-back and voice agents as some of the highest-ROI automations a local business can install. It's a different thing to see a venture firm underwrite that exact thesis at a billion-dollar valuation. So this post is less "here's a new automation idea" and more: here's what a $125 million bet on a ringing phone tells us, and what it means if your business has five trucks instead of five thousand customers.
Coverage of Avoca's raise cited industry figures that are worth sitting with. HVAC contractors reportedly lose an average of around $45,600 a year to missed calls; for plumbers, that figure climbs toward $125,000. About 74% of contractor calls are said to go completely unanswered. The average missed call is estimated to cost roughly $1,200 in lost revenue, and after-hours emergency calls, burst pipes, dead HVAC units in July, flooding, tend to run $450 to $600 per ticket versus around $275 for a standard scheduled visit.
Read those numbers as directional rather than exact, they'll vary by market, trade, and call volume, but the shape of the problem is the part that matters: the calls a trades business is most likely to miss are disproportionately the ones worth the most money. A missed call at 11pm on a Saturday isn't a lost $50 tune-up. It's a lost $8,000 emergency system replacement that goes to whichever competitor actually picks up.
Avoca's product is narrow on purpose: an AI voice agent that answers every call a home services business can't get to, captures the caller's name, number, and job details, and books it straight into the scheduling system. Reporting on the raise put the lift in lead-to-booked-job conversion at 25 to 40% within the first 90 days for most operators, driven almost entirely by the fact that overflow and after-hours calls stop going to voicemail. The company's customer list includes 1-800-GOT-JUNK?, Goettl, and Turnpoint Services, and it's reportedly on track to help customers book around $1 billion in jobs in 2026.
None of this is exotic engineering. It's a voice model, a scheduling integration, and a very tightly scoped job: answer, qualify, book, hand off anything unusual to a human. That narrowness is exactly why it works reliably enough for a home services company to trust it with real customers, and it's the same principle I've written about before when picking apart missed-call text-back for home services: the automations that make money are boring, single-purpose, and easy to audit.
Here's the part that matters if you run a five-truck HVAC company instead of a national franchise chain: the underlying stack Avoca is charging enterprise prices for is buildable at a fraction of the cost. A voice platform like Retell AI or Vapi handles the call itself, an orchestration layer like n8n routes the captured details into your scheduling tool and CRM, and a text-back fallback catches anything the voice agent can't resolve. I've broken down the platform choice in more detail in Retell AI vs Vapi vs Bland, and the running costs in the 2026 voice agent cost breakdown.
The difference between a $1 billion platform and a freelancer-built version isn't the core mechanic, it's scale, support infrastructure, integrations with every major field-service CRM out of the box, and a sales team. If you're a single-location trades business, you don't need any of that. You need the phone answered and the job booked.
A practical build for a trades company with a handful of trucks usually looks like this:
That's four components, not forty. Every one of them is off-the-shelf technology stitched together with clear rules about what the system is and isn't allowed to decide on its own.
Avoca's customer base spans HVAC, plumbing, roofing, and electrical, but the pattern holds for any trade where jobs are won by whoever responds first and where emergencies happen outside business hours: pest control, garage door repair, appliance repair, locksmiths. Reporting on a pest control operator using a comparable AI answering setup noted two consecutive weeks of zero missed Google Local Services calls, which matters specifically because those are high-intent, pay-per-lead calls that are expensive to generate and painful to waste.
The common thread across all these trades is the same: emergency and after-hours calls are rare relative to total call volume, but they're worth a disproportionate share of annual revenue, and they're exactly the calls a two-person office can't staff for around the clock.
For a single-location trades business, a voice agent plus text-back fallback typically runs a few hundred dollars a month all in, covering the voice platform's usage fees, the orchestration layer, and phone number costs. Setup is a one-time project cost that scales with how many call types and edge cases you want handled well from day one. Compare that to even one recovered $8,000 emergency job a quarter, and the math resembles the bet Avoca's investors made, just at a scale that fits a business with a handful of trucks instead of a national fleet.
The HVAC industry alone was estimated at roughly $50 billion in 2025, with some projections putting it near $75 billion by 2032. Avoca didn't get funded because answering the phone is a novel idea. It got funded because the market underneath that idea is enormous and the automation still isn't standard practice for most operators.
A $1 billion valuation for "answer the phone better" is a strong signal, not because it means every trades business needs the same platform, but because it confirms the economics were never the risky part. The risky part was always execution: building something reliable enough that a business owner trusts it with a customer's first impression. That's solvable at small-business scale with the right scope and the right guardrails, and it doesn't require a Series B to prove it works for your market.
If you're not ready for a full voice agent, start with missed-call text-back, it's cheaper, simpler, and still recovers a meaningful share of what you're currently losing. Move to a voice agent once you're consistently missing calls during business hours or want genuine 24/7 coverage for the emergencies that pay the most.
Gideon Wafula builds custom AI voice agents, missed-call text-back, and n8n automation systems for local service businesses.
See Services →