At its annual WebexOne conference last week, Cisco unveiled additional artificial intelligence (AI) features in its Webex suite that build on those it announced in March, all of which aim, it said, to improve both hybrid work and customer experience.
It was, said executive vice president, security and collaboration Jeetu Patel, during his keynote, “probably the most consequential set of innovations that we have done in the history of Webex.”
He also noted that the event was special for a second reason: “A few years ago, there was a lot of work that we needed to do. And a lot of you actually bet on us, and believed in us, and gave us feedback, and gave us the input, and pushed us to get better. And this event is actually for all of you, to make sure that we can express our gratitude, and we can make you proud for the decision that you’ve made.”
The company’s two primary initiatives today, he said, are hybrid work and customer experience. “I think there’s going to be a fair amount of transformation that needs to happen so that hybrid work actually works. Because today, it doesn’t really work that well. And customer experience is pretty broken, too.”
The new AI tools will address both areas, he said, adding that in fact, “at this event, what you’re going to hear is that AI is pervasive in every aspect of the Webex platform.”
Patel’s first announcement was the Real-Time Media Model (RMM) that complements large language models (LLMs) to add context to meetings. “The models have the ability to take multiple media streams and produce multiple outputs, such as people and object recognition and action analytics like movement and gestures. RMMs in Webex will also enable audio and video channels to be used as signals of context in traditional text-based capabilities like meeting summaries and highlights,” Cisco said in a release.
To help hybrid workers connecting in over low bandwidth, a new AI-driven audio codec will produce high-quality sound even over low bandwidth connections, and Super Resolution video will do the same for the visuals.
And, of course, there’s an AI assistant that will operate across all components of the suite. It is already embedded in Webex devices. It will allow users to catch up on missed meetings, or missed portions of meetings in real time. There’s also a collection of upcoming capabilities, including:
- Change Message Tone, which provides recommended changes to tone, format, phrasing in Webex Messaging and Slido. For example, the Webex AI Assistant may suggest changing a blunt response like ‘this isn’t good; rework the summary,’ to a more professional option of ‘this is a fantastic document, but the upfront summary could use some improvement.’
- Suggested Responses to messages for contact centre agents responding to customers on digital channels. Suggestions of text are informed by context history and surrounding intelligence.
- Meeting Summaries will help users catch up on missed portions of meetings or entire meetings in real-time, with an easily digestible summary of what was missed. They will be organized into chapters and highlights. With chapters, people can skip to various topics. For example, they may prefer to focus in-depth on the written portion of the finance discussion, then watch a brief video that summarizes the focus of the meeting in its entirety.
- Message Summaries help people stay up to date by recapping unread messages or recapping spaces. They are customized, so a user can ask Webex AI Assistant to summarize messages from their boss in a space that contains messages from multiple people, for example.
- Slido Topic Summaries generate common topics shared in interactive Q&A and polling during virtual and hybrid events for participants to easily navigate what’s trending.
These new capabilities are planned to begin appearing in the products before the end of 2023.
The company also announced that it will use a combination of large language models, using commercial, open source, Cisco proprietary, and customer models, as appropriate, for its AI assistant.
“We wanted to make sure that we gave you your choice,” Patel said. “Whether it’s a commercial model, whether it’s our own model, or whether it’s a select set of customer models for things that we need to make sure that we can actually enable for certain industries, for certain vocabulary, we will make sure that those are made available with an orchestration layer that has access to that from the Webex platform.”
The orchestration layer will not only help customers direct their prompts to the appropriate model, he noted, but will provide a single point of policy enforcement to maintain data privacy and security.
With the new products and features created with responsible AI principles in mind, he said “our commitment to you is the following. Your data belongs to you, not to us. Number two, security is by design.
“Because I do two jobs at Cisco – one job is to run Webex, the second job is running security – security is by design in the core fabric of what we do. And number three, trust is earned every single day, but can be lost in an instant. So we are going to make sure that transparency is paramount.”
