The useful version of AJ Jain's founder story begins on the road to Sacramento. Jain and his co-founder, Aditya Ketkar, were commuting from the Bay Area to spend time with plumbing contractors. The daily round trip ran to four hours. It was inefficient in exactly the way early customer research can be efficient: the distance forced them to stay, watch and listen.
They went door to door and spoke with more than 100 contractors. Inside those businesses, the vocabulary changed. There were dispatch boards, on-call technicians, open estimates and expensive leads arriving after the office had closed. Owners were not asking for an abstract AI strategy. They wanted the phone answered. They wanted the customer understood. They wanted the job booked in the software their teams already used.
Jain has described the method in practical terms: gather enough firsthand detail to see the pattern, narrow it to the three recurring problems, then work on the one with the largest impact. The ringing phone won. For a plumbing or HVAC company, a missed call is rarely just a communication failure. It can mean a lost emergency job, wasted marketing spend and a customer who immediately dials the next result.
Before the plumber, the electric vehicle
Broccoli was not Jain and Ketkar's first company, or even their first attempt together. The pair met in a New Venture Creation course at BITS Pilani's Goa campus, where both studied computer science. Their working relationship outlasted the class. Ketkar later worked at Microsoft; Jain joined Adobe, first as an intern and then as a software engineer. By 2021 they were building Telematica.
Telematica was an API platform for electric vehicles. It pulled telematics, battery and mobility data into a standardized interface, and the founders also built energy-management systems for European utilities and charging-point operators. The company entered Y Combinator's Winter 2022 batch. It raised $1.5 million. It had customers and a credible technical problem.
But a company can work without becoming the company its founders believe they should spend years building. Telematica was growing, yet the pair did not see the scale they wanted. With support from Y Combinator, they pivoted. That decision is easy to flatten into a clean arrow on a timeline. In practice, it meant walking away from accumulated knowledge, entering an unfamiliar industry and returning to beginner status.
The pivot worked because the founders did not bring a fixed product to the trades and ask owners to admire it. They treated proximity as a form of research. The contractors supplied the nouns. The engineers supplied the system.
The technology earned its place only after the workflow had spoken.The field lesson in one sentence
An AI product measured in booked jobs
Broccoli's first wedge was an AI voice agent for calls that human teams could not reach: nights, weekends, sudden spikes and overflow. It would follow the contractor's script, book into ServiceTitan or another CRM and trigger an escalation when the job was urgent. From there, the product expanded across the front office. Today Broccoli markets agents for reception, sales, appointment reminders, membership renewals, open estimates and post-job “happy calls.”
The product names are deliberately ordinary. Dane is a customer-service agent. Amy is a receptionist. Mark checks whether a customer is happy. Sean sells services. Ashley follows up on estimates. The friendliness is part of the interface, but the more consequential work happens behind the voice: turning a conversation into structured data, finding capacity, assigning the right workflow and keeping the office informed.
Broccoli presents an 85 percent-plus job-booking rate, a 22 percent increase in revenue and a two-day onboarding time. The figures are useful as a picture of what it chooses to optimize. The metrics are commercial and legible to an owner. There is no benchmark score in the pitch. There is the percentage of calls converted into work.
The vertical-AI stack, from voice to action
Jain's public case for AI in the trades is careful about the handoff between people and software. Contractors still want relationships. Customers still want a person when a situation becomes complicated. His argument is that responsiveness comes first: the company cannot create a human relationship with a caller it never answers. The agent covers the gap, records the interaction and routes the work.
This is why integration matters more than a convincing voice demo. A pleasant conversation that never reaches the dispatch board creates a new island of information. Broccoli's bet is that the agent should work inside the existing operating system of the contractor. It should know the schedule, log the transcript, notify the manager and leave a clean record for the morning team.
The customer who crossed the table
The most telling early relationship was with Marcelle Flowers of Armstrong Plumbing, a company with three decades of operating history. Flowers began as a Broccoli customer while the product was still learning the industry. Armstrong later saw weekend bookings rise 900 percent after using the platform. The number drew attention. Her next move said more: she joined Broccoli as Executive Industry Advisor and chief operating officer.
A customer becoming an operator inside the vendor is a particular kind of validation. Flowers brought the memory of what fails at 7 p.m. on a Saturday, how a service manager reads a dispatch board and what customers will tolerate when a pipe is leaking. Jain and Ketkar had technical depth. Flowers had the lived workflow. Broccoli could now place both views in the same meeting.
The company grew beyond its original wedge. Broccoli became an official ServiceTitan technology partner in 2025. It reached more than 800 contractors in under two years and a team of 85 by August 2026, up from the handful visible a year earlier. In July, the company announced a $25 million Series A led by Khosla Ventures.
The capital arrived quickly. Jain and Ketkar closed the round after 14 days of continuous follow-up. Fundraising speed makes a crisp headline, but the longer preparation happened elsewhere: nine years of co-founder history, one previous company, more than 100 contractor conversations and the operational evidence of customers such as Armstrong.
A village in purple polos
Jain's public writing is less formal than his title. Team members become “BroccHeads.” Posts arrive with vegetable emoji, trade-show snapshots and short bursts of enthusiasm. When the sales team gathered in San Francisco for dinners and a Giants game, his summary was eight words: “Absolute killers. Love this team!!”
A trade-show story reveals more about the culture he wants to describe. A colleague, Ryan, texted to ask whether he could bring extra help to set up the booth. The helpers were his children, on summer break. Jain laughed at the formality and described Broccoli as an “ask forgiveness not permission company.” He called the spouses, children and families around the team a village three or four times larger than the headcount.
It is a cheerful story, but it also fits the business. Broccoli sells into an industry where family companies are common, children visit the shop and an owner's identity can be inseparable from the work. The startup has learned to show up in that world: at trade shows, contractor podcasts and service-business conferences. Jain speaks about AI, then translates it into the lost lead after 5 p.m.
His path has been a series of translations. At Telematica, raw vehicle data became an API. At Broccoli, an anxious phone call becomes a booked appointment and an actionable record. In both cases the engineering disappears beneath an operational result.
A service business cannot book the job it never answers.The commercial idea at the center of Broccoli AI
The next call
Jain is now betting on something larger than an after-hours receptionist. He describes an AI operating layer for home services, spanning revenue and operations: inbound calls, outbound follow-up, quality review, coaching, scheduling, dispatch and reporting. Broccoli's recent hiring points in the same direction, with roles across engineering, product, partnerships and customer success.
The risk is familiar to any vertical-software company. Every new workflow adds local exceptions. A plumbing emergency is not an HVAC tune-up. A one-location shop does not operate like a 23-location group. The product has to scale without becoming generic, and automate without losing the moments when judgment or empathy should pass to a person.
Jain's early answer remains his most useful one: go back to the shop. Listen until the pattern is clear. Pick the problem with the largest consequence. Broccoli's story can be told with a $25 million financing and hundreds of customers, but its working method still fits on a field note.
The phone rings. Someone answers. The job moves forward. For the contractor, that is the product. For Jain, it is proof that the long commute led somewhere.