Why Georgian Invested in PolyAI (Again)

We are pleased to announce our continued partnership with PolyAI, co-leading their $86 million Series D funding round alongside Khosla Ventures and Hedosophia, with participation from NVIDIA’s NVentures and other returning investors.

PolyAI is a leader in the voice assistant market, building lifelike customer experience agents that are designed to handle over 95% of all customer interactions from start to finish without involving humans. PolyAI leverages proprietary, pre-trained machine learning models to deploy automated and enterprise-ready AI agents in weeks compared to months.

Since Georgian led the Series B in 2022, PolyAI has evolved from a technology pioneer into a valued infrastructure provider for global enterprises like FedEx, Marriott, PG&E and UniCredit. This latest round of funding brings PolyAI’s total raised capital to over $200 million and reaffirms our belief that PolyAI is creating a leading enterprise solution in customer experience automation.

The Origin Story: Rethinking Our Priors

We first invested in PolyAI in early 2022. At the time, we’d had a thesis on conversational AI for more than five years but had made only one investment in the space (Chorus). Despite our conviction that voice agents for customer support and sales represented a massive opportunity, we weren’t convinced by the product differentiation between the many solutions we saw entering the market. 

In early 2022 we met PolyAI’s CEO Nikola Mrkšić via an introduction from a co-investor in another Georgian portfolio company. As we became acquainted  with the PolyAI team, we reassessed our assumptions about the ability for a company to demonstrate product differentiation in the AI assistant market.

During Georgian’s initial diligence on PolyAI, our investment team conducted extensive market reference calls with customers and industry participants, and assessed PolyAI using an internal conversational AI evaluation framework. Throughout our conversations we were particularly impressed with PolyAI’s ability to match the voice quality of their AI assistants to a customer’s brand identity.

PolyAI’s product performance was backed by a deep technical pedigree; founders Nikola Mrkšić, Shawn Wen and Pei-Hao (Eddy) Su completed their PhDs at Cambridge, researching dialogue systems under the supervision of voice AI pioneer Steven Young. We felt fortunate to partner with them and lead PolyAI’s Series B round in May of 2022.

The Market Opportunity: Validation at Scale

Customer interactions represent one of the largest (and highest-ROI) opportunities for applied generative AI. McKinsey estimates that generative AI could unlock hundreds of billions of dollars in annual value in customer-facing functions such as customer operations, marketing and sales. 

We believe that realizing that opportunity will require omnichannel automation (including voice) to meet customers where they are. Despite the rise in text- and chat-based channels for customer engagement, voice remains an important channel for complex, urgent, and high-emotion customer needs including escalations. 

The ability for voice and chat agents to engage customers immediately, at scale, and with increasingly personalized experiences can be a significant loyalty, retention, and revenue driver for enterprises. AI agents can also generate meaningful cost savings by increasing containment, improving routing and resolution, and reducing the burden on human teams.

PolyAI has moved beyond theoretical potential to show tangible economic impact for its customers. According to a 2025 Forrester Consulting study, the PolyAI customers surveyed achieved a 391% return on investment, realizing an average of $10.3 million in cost savings. PolyAI’s platform drives this economic value via its omnichannel capabilities, with a particular emphasis on voice.

Why PolyAI? 

Over the past three years we’ve seen Poly scale from single-digit ARR to become a significant player in the voice AI category. Our rationale for re-investing in PolyAI includes:

1. Voice Quality, Team, and Platform

We believe that PolyAI is unique in the market due to the quality of their voice agents, the calibre of their machine learning team and the agent platform they have built.

PolyAI has built a proprietary stack specifically for spoken conversation. Their platform, Agent Studio, incorporates frontier LLMs alongside their own proprietary foundation models. This allows them to handle complex, multi-turn customer interactions in regulated industries where accuracy is important. Two examples include:

• Crisis Management at Scale (PG&E): During critical weather outages, Pacific Gas and Electric deployed PolyAI’s agents to handle emergency calls in English and Spanish. By automating these high-stress interactions, PG&E saved 35,000 labor hours and increased their Customer Satisfaction (CSAT) score by 22%.

• Modernizing Legacy Infrastructure (UniCredit): To replace a 20-year-old Interactive Voice Response (IVR) system—the traditional keypad menus that frustrate customers—UniCredit deployed a Croatian-speaking assistant. The new system routed calls 83% faster and boosted their Net Promoter Score (NPS) by 14 points.

2. Deep Workflow Experience

PolyAI has built differentiation not only in their technology stack but in how they make products work within customer environments. Their Forward Deployed Engineering (FDE) teams collaborate closely with clients to rethink workflows and integrate deeply with existing customer data systems. The goal is to  ensure that deployments aren’t just surface-level “chatbots” but core infrastructure designed to exceed customer ROI expectations. 

Looking Ahead

In the early days of Georgian’s Conversational AI thesis, our broader market concern was that “voice” would become a commodity wrapped around open-source models. Our experience working with PolyAI over the past three years—amidst a rapidly changing technology landscape—has demonstrated that PolyAI’s decision to build a full-stack, verticalized solution creates a defensible moat that generalist models cannot easily bridge.

Today, our conviction in PolyAI has been validated by nearly $1 billion in total value created annually for their customers. By continuing to invest in PolyAI we believe we are partnering with the infrastructure layer that will define the next decade of customer service. We are proud to continue building with Nikola, Shawn, Eddy and the entire PolyAI team as they transform the enterprise voice AI landscape.