Bland.ai's Bet on the Full Voice AI Stack

Voice AI is one of the more interesting categories in enterprise software and, one of the more crowded. Over the past year, we've spent meaningful time as a team getting to know companies all across this space. We met with model providers, infrastructure players, and orchestration layers. We came away convinced that voice will support multiple several-billion-dollar companies. We also came away with a clear view of what separates a durable company from a temporary one.

That view led us to Isaiah Granet, Sobhan Nejad and Bland.ai. Today the company announced a $50 million Series C, which we were proud to lead. Here's why we invested.

The Contrarian Decision That Mattered Most

Many voice AI products are wrappers. They sit on top of third-party foundation models and add a layer of orchestration. That approach is fast to build and easy to commoditize. But it also makes a product easy to leave when something better or less expensive comes along.

Isaiah and Sobhan made the much harder choice to own the full stack and that it be purpose-built for voice. We're betting this will prove to be the key long term differentiator in the space. When you control your own models, you can directly improve on what makes voice challenging: latency, ambiguity, continuity. And then from that foundation, build real solutions for the complex needs of the enterprise.

Built for Enterprise, Not Experiments

Owning the stack also lets Bland meet the bar that serious enterprises require. The platform is built with security, compliance, scalability, auditability, and deterministic reliability as first-class requirements, alongside the customer experience itself.

This matters because Bland's traction is concentrated in highly regulated industries, including healthcare and financial services. These are settings where conversations are long, complex, and consequential, and where a mistake carries real cost. The company now handles more than 3.5 million calls each week in those high-stakes environments.

We spent time with several of Bland's customers during diligence. The message was consistent. For them, voice AI is no longer experimental. It has become core infrastructure. One customer described adoption as not just essential but existential, with the view that organizations that fail to adopt will fall behind. We also heard that switching away from Bland is genuinely difficult once the platform is embedded, a reflection of that full-stack depth rather than marketing lock-in.

Voice Will Still Matter, Even as Agents Talk to Agents

A fair question is whether voice gets leapfrogged. As agent-to-agent workflows mature, some interactions will resolve without a person ever speaking. We expect that to happen, and for certain use cases it will be the right outcome.

But we don't believe it erases the value of voice. Many interactions still call for the human touch, and the way people connect with and experience a brand will continue to run through conversation. As models grow better at understanding context and responding with empathy, voice becomes more useful, not less. We see it settling in as a piece of core infrastructure for customer-facing operations across verticals, from support to scheduling to outreach.

Why Now and Why This Team

We back companies on the strength of their thesis, their technical depth, and the people carrying both forward. Bland gave us all three. Isaiah and the team struck us as a serious group solving one of the harder problems in AI, and the traction reflects it. We're confident that we can support them in turning that work into rapid, durable growth.

Here's to partnering with Isaiah and Sobhan and the entire Bland.ai team as they expand their research, grow their engineering team, and continue scaling out across the enterprise. The hard problems in voice are the ones worth solving, and this is a team built to solve them.

‍