YC Root Access

Rebuilding Customer Support for the AI Era

In this episode of Founder Firesides, YC Managing Partner Harj Taggar talks to the founders of Pylon (W23), Marty, Advith, and Robert, who recently raised a $51M Series B. Pylon is an AI-native customer support platform built for B2B companies, replacing legacy tools like Zendesk with a modern system designed for Slack, Teams, and other shared channels. It helps support, success, and account teams manage customer conversations, structure data, and automate workflows — without losing the high-touch relationships B2B requires. https://usepylon.com/ Apply to Y Combinator: https://www.ycombinator.com/apply Work at a startup: https://www.ycombinator.com/jobs

HTHarj TaggarhostADAdvithguestMAMartyguestRORobertguest

Feb 24, 2026·1h 0m·Watch on YouTube ↗

📖 CHAPTERS

1. 1

   What Pylon is today: AI-native customer support for B2B

   Harj introduces the Pylon founders and frames the conversation around their unusual path to finding the right idea. The team explains Pylon’s mission: an AI-native customer support platform built specifically for B2B companies, positioned to replace incumbents like Zendesk and Intercom.

2. 2

   Scale, customers, and funding: from YC W23 to Series B

   The founders share concrete traction metrics and the company’s current scale. They also name notable customers and outline their fundraising history through Series B.

3. 3

   Founder backstories: how startups became the inevitable path

   Each founder explains what pulled them toward startups, from early exposure and curiosity to dissatisfaction with large-company work. Their experiences (hackathons, internships, early startup roles) built the conviction—and working rhythm—that later enabled rapid iteration.

4. 4

   Starting while employed: routines, COVID leverage, and discovery calls

   They describe what day-to-day looks like when exploring startup ideas with full-time jobs. COVID-era remote work made it easier to schedule calls, and they progressively filled their calendars with customer discovery interviews.

5. 5

   First pivot mistakes: building too early and choosing hard markets (edtech)

   Their earliest concept was an alumni portal for student–alumni career connections. They reflect on common early-stage mistakes: building before validating, and tackling markets with difficult buying cycles.

6. 6

   Marty’s parallel path: many side projects and the decision to quit Airbnb

   Marty recounts exploring many ideas with different would-be cofounders and domains. He explains how dissatisfaction at work, combined with a “best-case/worst-case” analysis, made leaving feel rational and necessary.

7. 7

   Forming the three-founder team: cofounders over ideas (and early miscommunication)

   The founders describe how they eventually converged into a trio—after initial hesitation and even a soft rejection. They extract lessons about cofounder selection, trust, and why “idea-based” cofounding can block necessary pivots.

8. 8

   Choosing an idea systematically: market size math, public comps, and the ‘why now’

   They introduce a more analytical framework for evaluating ideas: work backwards from an ambitious outcome, analyze market size bottoms-up, and study why previous winners won. A key insight: horizontal B2B markets offer far larger ceilings than many “interesting” vertical ideas.

9. 9

   The spark: B2B customer support moving to shared Slack channels

   The breakthrough came from a friend describing a new operational reality: companies supporting customers directly in Slack. The team validated quickly across their network, recognizing the behavior as a broader post-pandemic shift in how businesses communicate.

10. 10

    High-velocity discovery: LinkedIn outreach, call structure, and daily iteration

    They detail their outbound discovery machine: daily personalized LinkedIn messages and high-volume interviews. The team shares practical call tactics—spending most of the time probing a hypothesis while leaving space for unexpected insights—and emphasizes how quitting jobs sped up iteration cycles.

11. 11

    Validation whiplash: Hightouch wanted it—and warned them not to build it

    A striking moment: Hightouch confirmed urgent demand immediately, yet their founder advised against pursuing the idea because they had tried it before. The team explains why prior failures aren’t disqualifying and how timing and execution can change outcomes.

12. 12

    YC entry and the first product: a narrow integration wedge and a deadline-driven sprint

    They describe an unconventional YC path: pre-idea office hours led to a deferred interview, which they used as a forcing function to close an initial customer. The first product was deliberately narrow: a Slack-to-Zendesk/Intercom ticketing integration built quickly to serve real demand.

13. 13

    From integration to platform: product expansion, AI-native support, and the road to CRM

    The team explains how customer pull and incumbent stagnation drove them from an integration into a full B2B support platform. They discuss AI adoption as workflow-led (not hype-led) and outline a future where Pylon evolves beyond ticketing into a customer-facing system of record—closer to “AI Salesforce” than “AI Zendesk.”