The real test of AI agents isn't how well they chat — it's whether they can automate the messy, offline workflows that still run most of the economy.
The Summary
- Ciridae raised $20M seed led by Accel to build AI agents for "real economy" businesses like HVAC companies, plumbers, and electricians
- Founders are ex-Apple and a16z operators targeting the 90% of businesses that don't have engineering teams or API integrations
- The pitch: AI agents that handle scheduling, dispatch, invoicing, and customer communication for industries still running on phone calls and spreadsheets
The Signal
Ciridae is betting on a market thesis most AI companies ignore: the real economy doesn't run on APIs. It runs on phone calls, paper invoices, and technicians driving to job sites. Co-founder Sarah Chen, who spent six years at Apple building enterprise tools, watched this gap firsthand. "Every AI demo I see assumes clean data, structured workflows, and an engineering team on the other end," she told Fortune. "That's maybe 10% of businesses."
The company is starting with home services — HVAC, plumbing, electrical — where the average business has five to twenty employees, zero developers, and workflows that look the same as they did in 1995. Ciridae's agents handle the full loop: answer inbound calls, check technician availability, schedule appointments, send invoices, and follow up on payment. No API required. The agent listens to how the business already operates, then mimics it.
"The average plumber doesn't need an AI coding assistant. They need an AI office manager who works 24/7 and never calls in sick."
Key traction signals:
- Closed $20M seed from Accel, a16z, and General Catalyst before public launch
- Beta customers report 40% reduction in missed appointments and 3x faster payment collection
- Founders include Chen (Apple), Marcus Liu (a16z growth team), and Tom Reeves (ex-ServiceTitan)
The timing matters. Agent frameworks like LangChain and AutoGPT made it cheap to build vertical AI agents, but most startups went after white-collar work: sales emails, meeting notes, data analysis. Ciridae is going the opposite direction. They're targeting industries where labor shortages are acute, margins are thin, and any efficiency gain drops straight to the bottom line. A plumber who can take twice as many calls without hiring a dispatcher just got a raise.
This isn't a consumer play. It's B2B infrastructure for the businesses that fix your toilet, install your AC, and rewire your house. Unsexy, high-margin, and massive. According to the U.S. Census Bureau, there are 33 million small businesses in America. Most of them look nothing like a tech company. Ciridae is building for them.
The Implication
The next wave of AI agents won't be trained on GitHub repos or Slack transcripts. They'll be trained on dispatch logs, customer voicemails, and handwritten invoices. If Ciridae delivers, they'll prove that the biggest AI opportunity isn't replacing knowledge workers — it's augmenting the people who keep the physical world running.
Watch for copycats in adjacent verticals: landscaping, pest control, auto repair. The playbook is the same everywhere. And if this works, the "real economy" might leapfrog tech companies in agent adoption. They have less to unlearn.