Retell AI: The “AI BPO” putting a voice agent on every phone line

(6 minutes)

Hi 👋, this is the Today in AI Newsletter: The weekly newsletter bringing you one step closer to building your own startup. 

We analyze a cool, industry-shaping AI startup every week, with a full breakdown of what they do, how they make money, how much they’ve raised, and the opportunity ahead.

Let’s get to the good stuff in this email:

• 📲 This startup is building the AI BPO to replace call centers forever ($350B Market)

• 💸 They’re already at $15M ARR with a ~20-person team, processing 30M+ calls/month, and ranked #2 on the Harmonic Hot 25 entering Q4 2025.

• 💰 Raised $4.6M Seed led by Alt Capital, backed by Y Combinator (W24) and 20+ founders/executives (including Aaron Levie and Michael Seibel).

So what’s the startup and who are the founders behind it? Here’s the story of Retell AI 📈

Retell AI was founded in 2023 by Bing Wu, Zexia Zhang, Todd Li, Weijia Yu, and Evie Wang to build conversational voice AI. 🗣️

They’re currently the #1 platform for building and managing conversational voice AI - think contact center automation without the wait music, or months of engineering. 

Their mission: cut implementation for voice agents from months to a day, and deliver a full-stack “AI BPO” that handles tasks end-to-end.

What exactly it does 👇

Retell combines an AI-native platform with a robust API so teams can build, test, deploy, and monitor agents that sound natural and resolve real work. 

The flow is straightforward:

1. You build with an agent builder and voice AI API. 🔊

2. You test using LLM-driven evaluations to harden edge cases. 🤖

3. You deploy across phone, web, SMS, and chat. 💬

4. You monitor success rates, latency, sentiment, and failed calls in a single dashboard to keep quality tight. 🧑‍💻

The special sauce is a custom turn-taking model that listens for semantic and audio cues, preventing interruptions and awkward gaps so calls feel empathetic and human.

If you need a mental model, think of it as putting your best operator - one who never gets tired, never forgets policy, and can book, route, log, and follow up on every line, 24/7.

Backstory 👀

Day 0 wasn’t customer service. The team started with live-streaming dubbing to help creators reach English audiences, but a painful 7-second latency killed traction. ❌

Then came the pivot spark ✨: community feedback asked, “Why can’t this just be AI?” 

That question pulled the founders into true two-way conversation - where latency, turn-taking, and handoffs actually decide if users trust the experience.

They launched publicly early on (HN/Product Hunt buzz), were initially rejected by Y Combinator, then accepted to Y Combinator (W24) after perfecting their demo.

They then shifted from a developer-only API to contact-center-grade agents - aiming to be the AI BPO that glues into real operations.

The founding team is super cool:

• ✅ Bing Wu (CEO): Ex-ByteDance/TikTok PM shipping to billions, two investor-backed startups with six-figure revenues.
  

• ✅ Zexia Zhang (CTO): Google alum shipping next-gen speech translation & NLP for call analysis.
  

• ✅ Todd Li (President): Repeat founder, ex-Google Ads engineer.
  

• ✅ Weijia Yu (COO): Ex-Meta PM/Tech Lead; previously shipped full-stack AI (TTS/CV/chatbot) at ObEN (Series B).
  

• ✅ Evie Wang (CMO): Ex-ByteDance product designer; led onboarding that cut drop-offs by 25% and launched a flagship app.

They have a large founding team compared to most startups but this means deliberate tension: be fast when speed wins, be stubborn when quality matters.