Pylon, a three-year-old SaaS company with around 1,600 B2B customers, demonstrated at SaaStr AI Day how its AI-powered customer support system deflected 50% of support tickets without reducing its 1,000-person support team. The company launched this agentic customer support system the week before the event, showing significant improvements in efficiency without headcount changes, according to saastr.com.
Co-founders Marty Kausas, CEO, and Advith Chelikani, CTO, explained that traditional AI support tools focused on deflection rates, which did not translate to reduced workload or headcount. Instead, Pylon’s approach combines human agents with AI assistance, enabling support staff to handle tickets more effectively. They shared examples where AI pre-investigated tickets, including one that replaced a manual AWS log query. Beta results from customers showed a 70% reduction in escalations and a 64.5% faster first response time.
The significance of Pylon’s approach lies in challenging the prevailing industry focus on deflection metrics, which often fail to reduce support costs or improve customer experience. By augmenting rather than replacing human agents, Pylon’s system aligns with evolving support roles, where engineers manage AI-assisted agents rather than coding directly. This model contrasts with many AI deployments in customer support that have not yielded expected headcount reductions or efficiency gains.
Pylon’s next milestones include expanding adoption among its 1,600 customers and refining its AI tools to further improve support outcomes. The company’s co-founders emphasized that the true test of AI in support is whether the job becomes unrecognizable compared to past workflows. The beta results with 70% fewer escalations and faster responses mark concrete progress in this direction, as reported by saastr.com.