PolyAI scores $12M Series A to put its ‘conversational AI agents’ in contact centres

PolyAI, a London startup founded by experts in the field of “conversational AI” — including CEO Nikola Mrkšić who was previously the first engineer at Apple-acquired VocalIQ — has raised $12 million in Series A funding to deploy its tech in customer support contact centres.

The round was led by Point72 Ventures, with participation from Sands Capital Ventures, Amadeus Capital Partners, Passion Capital, and Entrepreneur First (EF). PolyAI’s founders are graduates of EF, although they didn’t meet during the company building program but already knew each other from their time at Cambridge’s Dialog Systems Group, part of the Machine Intelligence Lab at the University of Cambridge.

“We started PolyAI in 2017, straight after submitting our PhD theses,” Mrkšić tells me. “At Cambridge, we developed state-of-the-art conversational technology, and starting a company was the best way to get this tech used in the real world. We brought many of our Cambridge colleagues with us and started building the commercial version of our conversational platform”.

Targeting contact centres — in a bid to help make these low-margin businesses more scalable — PolyAI’s AI tech doesn’t just attempt to understand customer queries but ensure they can be conducted in a truly conversational way, regardless of the medium, which could be over email, messaging or voice. Where a lot of conversation AI or voice assistants fall down, says Mrkšić, is that they aren’t able to really follow a conversation, often lacking the ability to understand meaning within the context of a conversation’s history or follow-up dialogue.

“Our proprietary technology allows the AI agents to support really complex use cases,” he says. “Our agents are built around a framework for modelling context, which means they can hold long conversations and remember all pieces of information that users had previously shared. The backend models are data-driven, and they are domain and language agnostic. This allows them to seamlessly scale across different use cases and world languages. In practice, this means that we don’t have to hand-craft agent behaviour — AI agents can learn by observing human agents at work”.

That’s a hard nut to crack, which is why Mrkšić believes deep vertical integration with contact centres will produce the best outcomes. He doesn’t rule out either buying a small to medium-sized contact centre or forming a strategic partnership to expedite improvements in PolyAI’s offering and the company’s understanding of how contact centres operate. His thesis is that AI can help make contact centres more profitable, although, early on in the startup’s life, the case is not yet proven.

Related to this, Mrkšić and his team aren’t proposing that “AI agents” replace human agents altogether but work alongside them, quite literally, with each playing to their respective strengths. PolyAI co-founder and CTO Shawn Wen argues that machines can do many things that humans struggle with, including having “instant access” to all of relevant information needed to support a customer. At peak times, this can mean AI agents handling calls autonomously if human agents aren’t available, while leaving human agents with the more complex edge cases or ones where they can bring the most value through human empathy and EQ.

“We plan to pursue very tight integration with contact centres — be that through M&A, investment or other profit-sharing arrangements,” adds Mrkšić. “Whichever model we end up pursuing, we want full alignment between PolyAI and contact centres. Too many AI companies have died trying to find favourable software licensing agreements years before their technology was ready for wide-scale deployment. We believe vertical integration is the best way to fast-track the development of our ML platform, as well as for PolyAI to stay independent in the long-term”.

This post was originally posted at http://feedproxy.google.com/~r/Techcrunch/~3/I86AyTLGf8k/.

Leave a Reply

Your email address will not be published. Required fields are marked *