As generative AI has taken center stage in recent months, Zoom has been looking at partnerships with LLM companies as a hedge against an evolving AI landscape. Today, the company announced a new partnership with Anthropic to use its Claude chatbot on the platform, starting with the Zoom Contact Center.
This comes on the heels of a partnership with OpenAI that the company announced in March, as it looks for ways to build more generative AI functionality into the Zoom family of products.
Smita Hashim joined Zoom three months ago as chief product officer, just as ChatGPT was capturing everyone’s imagination, after more than two decades of product experience at Microsoft and Google. She says that the generative AI landscape is shifting so quickly that the company decided to take an open approach where they could plug in different models as the industry rapidly changes.
“So when we started on our generative AI journey, we made a very conscious decision to go with what we are calling the federated approach to AI. That means the way we are designing our products is that we can use different models in the back end,” Hashim told TechCrunch.
The Anthropic partnership is part of that, starting with a focus in the contact center. “In particular for Anthropic and Claude we want to start with our Contact Center software because of its Constitutional AI approach and being able to follow rules very closely,” she said.
It will be essential to provide more accurate answers to customers, while limiting hallucinations where the models make up answers when they don't know the correct response. “It’ll be helpful to have that kind of more prescriptive-module-based approach [with Anthropic’s constitutional AI]. So that's where we want to start our work, and start bringing some of that functionality to market.”
Zoom intends to keep working with the Anthropic model, and as they learn more about how it works, eventually use it in other Zoom products. It's important to note, however, that there is no set date to bring any of the Anthropic-based functionality to market yet, but the company will continue to work on it, she said.
While Hashim, wouldn't say if the company is working on its own LLM, she says that these partnerships are a starting point for the company. It will continue to explore other options, including emerging open source models, even as it works to incorporate the OpenAI and Anthropic models across the Zoom platform.
While it was at it, Zoom announced that its venture arm, Zoom Ventures, invested an undisclosed amount in Anthropic under the terms of the partnership, presumably to have some skin in the game and invest in the technology they intend to be using in the future.