PolyAI, a London start-up set up by experts in the field of "conversational AI" – including CEO Nikola Mrkšić who was previously the first engineer at Apple-acquired VocalIQ – raised $ 12 million in Series A financing for its technology in customer support contact centers to use.
The round was led by Point72 Ventures, with the participation of Sands Capital Ventures, Amadeus Capital Partners, Passion Capital and Entrepreneur First (EF). The founders of PolyAI are EF graduates, although they did not meet during the business program, but already knew each other in their time in Cambridge & # 39; s Dialog Systems Group, part of the Machine Intelligence Lab at the University of Cambridge.
"We started PolyAI in 2017, immediately after submitting our theses," says Mrkšić. "In Cambridge, we developed advanced conversational technology, and starting a business was the best way to use this technology in the real world." We have taken many of our Cambridge colleagues and started building the commercial version. from our conversational platform ".
Focus on contact centers – in an attempt to make these companies with a low margin more scalable – PolyAI's AI technology not only tries to understand customer questions, but also allows them to be conducted in a genuinely friendly manner, regardless of the medium, which can be via e-mail, messages or speech. Where a lot of conversation AI or voice assistants fall down, says Mrkšić, is that they can not really follow a conversation, often lacking the ability to understand meaning in the context of the history of a conversation or a follow-up dialogue.
"Thanks to our own technology, AI agents can really support complex usage cases," he says. "Our agents are built around a context modeling framework, meaning they can have long conversations and remember all the information that users had previously shared." Backend models are data-based and agnostic for domains and languages. they can scale seamlessly across different use cases and world languages - in practice this means that we do not have to perform the behavior of agents – AI agents can learn by observing human agents at work ".
That is a hard nut to crack, so Mrkšić believes that deep vertical integration with contact centers yields the best results. He does not exclude that he buys a small to medium-sized contact center or enters into a strategic partnership to accelerate improvements in PolyAI's offering and the company's insight into the operation of contact centers. His thesis is that AI can help make contact centers more profitable, although the case has not yet been proven in the beginning of the startup.
In this connection, Mrkšić and his team do not suggest that "AI agents" completely replace human agents, but literally work side by side, with each game for their respective strengths. PolyAI co-founder and CTO Shawn Wen states that machines can do many things that people struggle with, including having "direct access" to all relevant information needed to support a customer. At peak times, this may mean that AI agents handle calls autonomously when human agents are unavailable, while human agents have the more complex edge cases or where they can bring the most value through human empathy and EQ.
"We intend to pursue very close integration with contact centers – whether through mergers and acquisitions, investments or other profit sharing schemes," Mrkšić adds. "Whatever model we pursue, we want full alignment between PolyAI and contact centers, too many AI companies have died in an attempt to find years of favorable software licensing agreements before their technology was ready for large-scale deployment." We believe that vertical integration is the best way is to accelerate the development of our ML platform, as well as for PolyAI to remain independent in the long term ".