YC Root Access

This Startup Built AI That 80% of Callers Think Is Human

In this episode of Founder Firesides, YC General Partner Nicolas Dessaigne talks to Will Bodewes, founder and CEO of Phonely (S24), which just raised a $16M Series A led by Base10 Partners. Phonely is an AI-powered voice platform that answers phone calls for businesses — handling millions of calls a month across hundreds of verticals using custom LLMs that statistically optimize outcomes over time. They discuss how watching his dad struggle to answer phones sparked the idea, going from small business customers to enterprise call centers, and why by the end of this year most callers won't know they're talking to AI. 0:00 Intro & Series A announcement 0:18 What Phonely does (Voice AI platform) 0:43 Scale: Millions of calls per month 0:56 Customers: Call centers, insurance, home services 1:29 Voice AI boom & differentiation 1:55 Founder background & early PhD work 2:44 Optimization advantage (data-driven improvements) 3:12 Founder journey: athlete → AI PhD → startup 4:41 Origin story: dad’s business problem 5:06 Early product → iteration loop 5:49 Pivot to call centers (enterprise focus) 6:08 Building custom models vs using OpenAI 7:05 Architecture: modular models for latency & cost 7:54 Is latency still a bottleneck? 8:23 How good is AI on phone calls today? 9:01 Should AI disclose itself? (ethics + regulation) 10:10 Real-world use cases (revenue-driven inbound calls) 10:47 Inbound vs outbound strategy 11:13 Series A story 12:28 What’s next for Voice AI 12:59 Competition vs generic models 13:23 Future vision (50M+ calls/month) 13:49 Hiring (sales + engineering) 14:21 Founder lesson: reality vs perception 15:26 Advice for founders (who should start companies)

NDNicolas DessaignehostWBWill Bodewesguest

Apr 16, 2026·16m·Watch on YouTube ↗

📖 CHAPTERS

1. 1
   0:18 – 0:43

   What Phonely actually sells: outcome-optimized voice agents

   Will clarifies that “answering the phone with AI” is only the surface-level description. The deeper value is helping businesses measure and improve the agent’s performance on customer-defined outcomes.

2. 2
   0:43 – 0:56

   Proof of scale: millions of calls/month across many verticals

   Phonely operates at high volume and across a broad set of industries. The company highlights that it’s not a single-vertical solution, implying robustness in production deployments.

3. 3
   0:56 – 1:29

   Who buys: call centers, insurance, home services—and why they care

   The primary customers are organizations with call-center workflows where lead qualification, appointment booking, and conversion are revenue-critical. Phonely’s pitch is consistent performance plus measurable improvement over time.

4. 4
   1:29 – 3:12

   Differentiation in a crowded voice AI boom: production know-how + optimization

   In response to rising competition, Will argues Phonely’s edge comes from early work before “voice AI” was a category and from deep experience deploying in production. He also emphasizes building tooling and data loops that quantify what changes improve outcomes.

5. 5
   3:12 – 4:41

   Founder backstory: athlete, first startup, then an AI PhD in Australia

   Will explains a non-linear path: collegiate cross-country skiing disrupted by COVID, an early failed company, then a fully funded AI PhD in Melbourne. That period became the foundation for learning AI deeply and beginning voice-focused work.

6. 6
   4:41 – 5:06

   The origin spark: solving his dad’s “someone answer the phone” problem

   Phonely began as a direct response to a real small-business pain point: missed calls and limited capacity to answer phones. When Will couldn’t find existing software that solved it well, he started building it himself.

7. 7
   5:06 – 6:08

   Early GTM and iteration loop: start small, learn fast, then pivot upmarket

   The company initially targeted small businesses at low price points to get rapid feedback and shorten iteration cycles. Within months, landing a call center changed the economics and focus, driving a pivot toward enterprise/call-center customers.

8. 8
   6:08 – 7:05

   Why build custom models instead of just using OpenAI: latency, control, and cost

   Will explains that Phonely moved beyond off-the-shelf closed models, collaborating early with Groq to address inference speed. Their architecture favors smaller specialized models for different components of a call, improving latency and economics while retaining quality.

9. 9
   7:05 – 10:10

   Inside the architecture: modular components for memory, variables, and tasks

   Rather than routing by question type, Phonely decomposes the voice agent into functional modules (e.g., capturing/storing a name/email) that can be optimized independently. This modularity enables isolating weaknesses and updating specific components without retraining everything.

10. 10
    10:10 – 11:13

    Where adoption is happening: inbound, revenue-critical calls (not generic support)

    Phonely sees strongest uptake in inbound calls tied directly to revenue, such as leads from billboards and ads. The agent qualifies callers, books appointments, or hands off to humans when required (e.g., licensed insurance agents).

11. 11
    11:13 – 13:23

    Series A story: ultra-endurance cycling posts led to Base10 connection

    Will shares that a LinkedIn post about ultra-endurance cycling and founder grit sparked a conversation with Base10’s Caroline. After continued discussions, Base10 preemptively offered the Series A, leading to a relatively concentrated fundraising process.

12. 12
    13:23

    What’s next: core technical improvements, competition, growth to 50M+ calls, and hiring

    The closing segment covers the roadmap (interruptions, endpointing, transcription), competitive positioning versus generic models, and aggressive scale targets. Will also shares hiring plans in SF and ends with lessons about the grind of founding and advice on who should start companies.

13. 13

    Phonely’s $16M Series A and the “AI answers your phone” premise

    The conversation opens with Phonely’s $16M Series A led by Base10 and a quick framing of what the company is building. Will Bodewes introduces Phonely as more than an AI receptionist: it’s a platform focused on improving outcomes over time.

14. 14

    State of the art: latency is ‘good enough,’ quality and accuracy are the frontier

    Will argues latency used to be a major bottleneck but is now more like table stakes. The bigger remaining challenges are conversational quality and accuracy, especially under noisy phone conditions.

15. 15

    How human is it really? 80% don’t realize it’s AI, plus disclosure ethics

    Will claims that roughly 80% of callers don’t realize they’re speaking to AI today, with expectations that this approaches near-universal indistinguishability soon. The discussion then turns to whether AI should disclose itself, with stronger disclosure expectations for outbound calling and emerging regulation.