Building the operating system for B2B support - one Slack thread at a time.
Every enterprise software company eventually builds a support system. Most build it around a ticketing queue and a prayer. Advith Chelikani built Pylon - a company premised on the idea that the entire paradigm was wrong.
Pylon's thesis is simple enough to fit on a napkin: B2B customers don't want to open a ticket portal. They're already in Slack. They're on Microsoft Teams, Discord, email. The support experience should meet them there - with AI agents that handle the routine, and human agents who actually have context when they show up. Advith has been building and re-building that vision since co-founding Pylon in November 2022 alongside Marty Kausas (CEO) and Robert Eng.
The company raised a $17M Series A from Andreessen Horowitz in 14 days in August 2024 - the kind of close that makes investors uncomfortable with how little time they had to deliberate. A year later, the Series B brought in another $31M, co-led by a16z and Bain Capital Ventures. Total funding: $51.2M. Total enterprise customers: 750+. Year-over-year revenue growth: 5x, twice in a row.
None of this happened because someone wrote a particularly persuasive deck. It happened because Advith treats customer calls the way other founders treat investor calls - with urgency, preparation, and volume. He does 6-8 customer calls per day and personally handles support tickets. At a company with 120 employees and $51M raised, that's unusual. Advith treats it as the entire point.
Reframe the problem always and keep it framed that way as you're building.
- Advith Chelikani, CHURN FM PodcastThe problem Pylon is solving isn't "customer support is slow." It's closer to: enterprise support infrastructure was designed for a world before Slack existed, before AI could draft a response faster than a human could open a ticket, and before B2B customers expected the same real-time interaction they got from consumer apps. Legacy platforms like Zendesk, Intercom, and Salesforce Service Cloud weren't built for this era. Pylon was.
The product reflects that framing with unusual discipline. Pylon integrates with Slack Connect, Microsoft Teams, and Discord natively - not as bolt-on integrations, but as first-class support surfaces. AI agents handle routine volume and reduce ticket work by 50%. AI assistants give human agents instant context and suggested responses, making them work 3x faster. The result is a platform that doesn't ask enterprise teams to change where their customers already are.
The label "Zendesk killer" gets attached to startups carelessly. Pylon earned it. When Bain Capital Ventures published a post titled "Inside Pylon's Mission to Build the Zendesk Killer," they weren't being provocative - they were describing the product roadmap. And the evidence was already stacking up: 150+ companies migrating off legacy platforms, 750+ customers, a list that includes Together AI, Cognition, Temporal, and AssemblyAI - some of the most technically demanding enterprises in the industry.
The migration number matters. Moving your customer support infrastructure is painful, expensive, and disruptive. Companies don't do it for marginal improvements. They do it when a new platform is meaningfully better for how their customers actually behave - and increasingly, enterprise customers behave in Slack.
Pylon built with AI from day one - not as a feature bolt-on. AI agents handle routine tickets automatically. AI assistants draft responses, surface context, and help agents move 3x faster. The architecture assumes AI is in the critical path.
Slack Connect, Microsoft Teams, Discord, email, in-app chat - Pylon treats these as first-class support channels, not integrations. Enterprise customers don't need to leave their workflow to get support.
B2B support isn't just volume - it's multi-stakeholder accounts, SLA contracts, internal runbooks, and success metrics. Pylon's data model was purpose-built for these workflows from the start.
Advith's path to Pylon has the cleanest logic of any founder origin story: he spent four years at Samsara as a Senior Software Engineer and Team Lead, scaling systems at an enterprise company - and came away with a precise understanding of where enterprise support infrastructure failed. He didn't read about the problem. He lived it from the engineering side.
Before Samsara, Advith interned at Slack in the summer of 2017 - the same summer he was a Kleiner Perkins (KPCB) Engineering Fellow. That fellowship planted two seeds simultaneously: it was where he met future co-founder Marty Kausas, and it gave him early exposure to the kind of product that B2B teams were building their workflows around. His Caltech CS degree gave him the technical depth; Slack showed him the surface on which enterprise work was moving.
The third co-founder, Robert Eng, came from Caltech itself - a connection predating the company by almost a decade. When all three eventually converged on the same problem in 2022, the founding team had a rare combination: Caltech-level technical rigor, enterprise-scale engineering experience, and a specific product thesis shaped by years of using (and being frustrated by) the existing tools.
The Caltech years also produced a GitHub that is, frankly, more varied than most: a blockchain voting system called Votechain, an ML-based trading algorithm, Slack trivia and puzzle bots, a Netflix recommender system. The range isn't decorative - it reflects a builder who doesn't stay in one domain and considers that a feature, not a constraint.
What Advith carries from those years into Pylon's culture is a particular philosophy about product velocity: shipping numerous high-quality features creates a kind of momentum that outlasts any individual decision. Customers stay not because switching is painful but because the product keeps getting better faster than they expected. That compounds. The 5x revenue growth two years running is what compound momentum looks like at the P&L level.
"Slope trumps experience always." For technical roles without hard prerequisites, Advith prioritizes learning velocity and intellectual capability over years on a resume. He's looking for the trajectory, not the current position.
6-8 customer calls per day isn't a founder quirk - it's a deliberate strategy. Product pain points and market positioning emerge from direct conversation, not analytics dashboards. Advith extends this principle to personally handling support tickets.
His core advice: reframe the problem and hold that frame throughout the build. Most product failures happen because teams get excited about solutions before they've truly understood the problem. The reframe is the work.
Shipping numerous high-quality features creates momentum that keeps customers engaged, even when onboarding and support processes aren't fully mature.
- Advith ChelikaniThere's a coherent through-line in how Advith approaches the craft of building. Reframe the problem. Stay close to customers. Hire for learning speed. Ship at high velocity. Each of these isn't a slogan - it's a decision rule that produces different behavior at a hundred daily forks in the road.
The market timing thesis behind Pylon reflects the same discipline. Advith and his co-founders identified the convergence of three independent forces: the B2B adoption of omnichannel communication post-pandemic, the emergence of AI tools capable of genuine ticket automation, and the growing unwillingness of enterprise customers to use separate portal interfaces for support. None of those forces was new in 2022. But their simultaneous maturity created a window - and Pylon was built precisely for that window.
The a16z investment thesis described it as "building for the future of B2B customer interactions." The Bain Capital piece framed it as "conversational and agentic." Both are accurate. What neither quite captures is the pace: the team that builds like they still live in a windowless apartment, working like the window will close if they're not fast enough.
Key technologies in Pylon's infrastructure:
Advith's intern experience at Slack in 2017 directly shaped Pylon's Slack-first support philosophy - a detail that makes the "why now" story feel less like timing and more like inevitability.
His GitHub includes Votechain (a blockchain voting system), an ML-based trading algorithm, and Slack puzzle bots - a range that suggests someone who builds for the joy of it, not just for the resume line.
Pylon's $17M Series A was closed in 14 days. For context: most seed rounds take longer than that. The speed reflects the momentum Advith and the team had built before the round even opened.
Advith met co-founder Marty Kausas at the KPCB fellowship in 2017 and co-founder Robert Eng at Caltech - two friendships incubated for years before the company existed. The founding team didn't form by accident.
February 2025. Navigating Early-Stage Growth: Churn, Retention & Customer Success. Advith on product velocity, hiring for slope, and staying close to customers as the company scales.
October 2025. "Pylon: Reimagining B2B Customer Support" - Advith joins co-founders Marty Kausas and Robert Eng on the a16z AI podcast to discuss the full company vision and product philosophy.
August 2024. "Pylon lands $17M investment to build a full-service B2B customer service platform." The Series A coverage that put the Zendesk challenger narrative into mainstream tech media.