By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
YC Root AccessThe AI Agents Helping Home Services Book More Jobs
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
Great conversations run long. Their best ideas don't have to.
uListen is free on iPhone, iPad, Android, Chrome, and Claude.
See howAt a glance
WHAT IT’S REALLY ABOUT
Avoca builds AI agents to capture missed home-services revenue calls
- Avoca started with an AI phone-answering agent for home services because missed or mishandled calls directly translate into lost high-value jobs.
- The founders argue AI is “bigger than SaaS” because it can capture labor/operations and marketing spend (moving from ~1% software wallet share toward ~15%+).
- They frame the product as augmenting—not simply replacing—CSRs in a high-attrition role, with new jobs emerging around supervising and training AI agents.
- Customer obsession and on-site workflow immersion drove product design, with a strict bar that features aren’t “done” until specific customers are happy using them.
- Growth to eight figures was fueled by deep love from a few design partners, referrals/influencer networks in contractor communities, and later an enterprise + private-equity consolidation sales motion.
IDEAS WORTH REMEMBERING
5 ideasPick markets where one interaction is financially meaningful.
Avoca chose home services after exploring restaurants because a single call can be worth $20K–$30K and phone calls drive ~85% of revenue, making call-handling automation immediately ROI-positive.
AI enables a shift from “software budget” to “labor and ops budget.”
They contrast ServiceTitan-like spend (~0.5–2% of wallet) with customer care/inside sales (3–5%) and marketing (7–10%), arguing AI agents can expand addressable wallet share toward ~15%+.
Lead with revenue capture, not cost cutting.
Although headcount reductions can happen (e.g., 100 CSRs down to 30 via attrition/promotions), Avoca sells primarily on booking more jobs and converting more marketing leads into revenue.
High-attrition jobs are prime for AI augmentation and role redesign.
With near-100% two-year attrition in CSR roles, taking 60–70% “nuisance calls” off humans can increase retention and enable promotions into dispatch or “AI agent trainer/manager” roles.
Human-in-the-loop is a product design discipline, not a fallback.
Drawing from autonomous-vehicle operations, they emphasize deciding when humans should step in and providing the right context so handoffs are fast and effective.
WORDS WORTH SAVING
5 quotesAt Avoca, we are building the AI workforce for the physical economy, starting with home services.
— Apurva Shrivastava
In a normal, um, home service business, there's almost a close to 100% attrition rate over the course of two years.
— Tyson Chen
If the pain is big enough, the customer will try to find a solution to it even if it's not perfect.
— Tyson Chen
Our job is to go follow where there's customer love, where people are gonna pay us more, and that's gonna be the more exciting opportunity for us.
— Apurva Shrivastava
You never think about it, but in home services there's this influencer network where folks like a Josh Campbell from ResQair, they are like kinda super connectors.
— Tyson Chen
Ask ChatGPT about this episode
Get its key ideas, plus a summary of each chapter linked to that point in the video, right in your chat.
Add uListen AI to ChatGPT.
Install uListen AI in ChatGPT (opens in a new tab)Needs a ChatGPT account.Paste this prompt into a new chat.
Use uListen AI to give me this episode's key ideas and a chapter-by-chapter summary, with timestamps: https://www.youtube.com/watch?v=EHp1HgwCo9I
SPEAKERS
Garry Tan
hostCEO of Y Combinator and the episode host/interviewer.
Tyson Chen
guestCo-founder of Avoca, building AI agents that handle inbound calls for home services businesses.
Apurva Shrivastava
guestCo-founder of Avoca, focused on AI agent software for home services customer support and booking.
High quality AI-generated summary created from speaker-labeled transcript.