Pylon: Rebuilding B2B Customer Support for the AI Era

How Three YC Founders Are Replacing Zendesk with AI

Advith, Marty, and Robert — the co-founders of Pylon — went through YC’s Winter 2023 batch and have since built one of YC’s fastest-growing companies: an AI-native customer support platform specifically for B2B companies. At 79 people, 8-digit ARR, 5.35x year-over-year growth, and over 1,000 customers including Honeycomb, Linear, and Incident.io, Pylon is a live case study in what happens when you rebuild a category from scratch with AI at the core.

The wedge that opened the market: Pylon started with a specific observation — B2B companies were increasingly communicating with customers over shared Slack channels, but had no tools built for that workflow. “I have a hundred customers and I talk to them over Slack. I have an outage. How do I let all these people know? I have to literally go into every single channel and copy and paste some message.” That visceral pain point became their beachhead into the much larger B2B support market.

Why the timing was right: Three forces converged: Zendesk’s acquisition by private equity meant the incumbent product was stagnating. B2B communication was shifting to Slack and Teams. And AI was making it possible to structure conversational data in ways that were previously impossible. “All this unstructured data that used to be really hard to turn into structured workflows is now structurable. That has been a huge unlock for us.”

The skeptic’s advantage: Interestingly, the Pylon team was initially resistant to adding AI. Rather than chasing the “AI hype,” they approached it from first principles: what are the actual customer problems, and what’s the best solution? AI happened to be the answer for most of them — but the discipline of starting from problems rather than technology led to more thoughtful integration.

B2B is different from consumer support: “If a million-dollar customer asks you a question, you can never answer them wrong and you never want it to feel impersonal.” This insight drives Pylon’s hybrid approach: some workflows are fully agentic, most combine AI with human judgment, and relationship-building stays purely human.

5 Lessons from Pylon’s Path to Product-Market Fit

• Start with a wedge, expand into the platform — Pylon began as a Slack ticketing integration ($400K ARR), then customers pulled them into building the full support platform. The founders didn’t need to know the end state before starting
• Big markets matter more than clever ideas — The team explicitly modeled market sizes before committing, comparing their logistics idea (tens of millions) against horizontal B2B SaaS (orders of magnitude larger). Market selection caps your ceiling
• Customer discovery at volume creates pattern recognition — 120 LinkedIn messages per day across three founders, converting to 30+ discovery calls per week. The quantity of conversations let them spot emerging patterns others missed
• AI-native beats AI-bolted-on — Building from scratch with AI at the core (rather than retrofitting legacy software) lets you rethink workflows entirely. Pylon structures conversational data across support, success, and product teams in ways Zendesk never could
• From Zendesk replacement to Salesforce competitor — The long-term vision isn’t just AI customer support; it’s an AI-native B2B CRM built on conversational data, expanding from ticketing into customer success, account management, and cross-team intelligence

What Pylon Means for AI-Powered Business Operations

Pylon represents a pattern that will repeat across every SaaS category: a new entrant rebuilds the category from scratch with AI at the foundation, while incumbents (acquired by PE, products stagnating) struggle to retrofit. The insight that conversational data — once unstructurable — can now power intelligent workflows across entire customer-facing organizations points to a future where AI doesn’t just automate tickets but fundamentally reshapes how businesses understand and serve their customers.