Agentic AI changes online meetings platforms: Passing from a silent observer to an active participant

Online meetings were almost the same as their colleagues from the physical world.

With the introduction of generative AI platforms for online meetings, new insights, including voice transcription services, began to add. Platform sellers, including Microsoft teams, Google Meet and Cisco Webx, have constantly integrated capabilities that go beyond what personal physical meetings can provide.

Now in the newly emerging era of the agent AI, online meetings are ready to diverge even more with a new wave of innovation. Zoom recently announced His AI agent efforts, which are intended to create the paradigm, move from meetings to basic stages. Microsoft added Copilot actions Which can be integrated with its Microsoft Teams service, helping users in meetings to get an idea and contact other Microsoft services. Cisco is constantly expanding its AI Opportunities in WebexAnd announced an Webex AI agent at the end of 2024, which helps to introduce the contact center.

Another company that has been particularly active in space is ReachWhich is a little differentiated in that it is not directly related to any particular platform for online meetings suppliers. While OTTER first made its sign as a voice transcription service powered by AI, it added AI assistant called Otter Pilot, AI AIT Assistant and a series of meetings known as meetings as GenniS

Today, Otter is a step forward, with its raid into agency AI. While some of the agents of the AI ​​features that OTTER adds are not unique, it does at least one thing that is not yet part of any Ait Agentic AI meeting technology. Otter Ai is already integrating as an entity in a meeting that can actually answer by voice to requests.

AI external participant is no longer available through a chat window only; AI is already a living unit that is literally part of the meeting.

AI meetings of AI meetings: Beyond the silent observers

Over the last few years, AI meetings have been passive observers-the conversation transfer, the creation of summaries and the authorization of requests after collection. OTTER is now changing this dynamic with its AI meeting agent, which can actively participate in calling calls.

“The new AI meetings agent we are building will be able to help you with a real -time voice,” Sam Liang, CEO of Otter Ai, told Venturebeat. “During the meeting, you can say,” Hey get rid of it and ask it. “

In a live demonstration with Venturebeat, Liang showed how the agent could answer factual questions, provide summary of meetings, and even plan subsequent meetings-all through voice commands during an active conversation.

What makes this particularly powerful is the agent’s ability to connect to the company’s ecosystem of the company.

“This agent can become a domain expert,” Liang noted. “When you have a meeting, it has almost endless knowledge from the Internet, but this agent also has knowledge of your enterprise.”

Autonomous Agents: When AI holds the meeting

By taking a step further, Otter also launches an autonomous SDR agent (sales development representative), which can independently hold entire meetings without human intervention. This agent congratulates visitors to websites, holds demonstrations of products and schedules for subsequent meetings with sales representatives.

“We cannot hire a million human agents to answer questions, but we have built this Otter SDR agent that functions as a sales development representative that can greet any visitor and give them a live demonstration,” Liang said.

The use of chatbots and avatar -based systems is not new. Liang claims that what distinguishes the technology of his company from existing avatar -based solutions is his ability to conduct demonstrations of multimedia products in real time. The autonomous agent can share screens, demonstrate product characteristics and answer specific questions about functionality and pricing.

AI agency’s technical architecture for meetings

Agentic AI is overloaded and somewhat excessively in the industry today.

Functionally agent AI is associated with activating actions that can be done by combining multiple models with a tool like Langchain or by using the features of the features present in many models.

Liang has an even more nun’s definition of Agentic AI.

“Agents are generally a more complex AI system that can degrade a large and complex task into less tasks,” he said. “This can make reasoning and can do some planning to complete a task.”

Otter does not use Langchain, but has developed his own personalized technology specifically designed for the challenges of the voice environment with many speakers. Technical architecture combines both information about the extraction of public knowledge and its own information about the enterprise through personalized performance of RAG (extracting a generation). This enables the agent to understand the company specific information such as employees names, project terminology and internal abbreviations.

The future of the reconnaissance meeting is not only an agency AI

What agent AI offers on online meetings platforms is a powerful new set of organizational efficiency opportunities.

For decades, organizational efficiency experts have warned of the risks of lost time in the meetings. Modern platforms powered by AI change this risk. Last week, Zoom’s CTO told me that its goal is to move the technology from meetings to basic stages, where the production production is not just another meeting, but for executive things that will benefit the organization.

While the AI ​​agent can create work processes, it still benefits with regular AI assistants who are not actually agent. There will still be independent AI assistants and completely agent, and this is a nice thing, according to Anurag Dhyingra, SVP & GM, Enterprise Connectivity and Cisco Cooperation.

“While AI agents act as autonomous employees and AI assistants serve as prompted helpers, both offer benefits for raising productivity and enhancing overall cooperation,” Dchindra told Venturebeat. “It is not about choosing one over the other, but rather to use their combined strengths to create an environment in which teams can focus on innovation and strategic decision making.”

What is also beginning to happen is more operative compatibility in different platforms. For example, Cisco’s AI assistant will work with work process applications such as Salesforce, Servicenow and Outlook.

Strategic consequences of agent AI for meetings of enterprises

For businesses evaluating their AI adoption road card, OTTER’s approach to agent meeting is a strategic folding point.

Instead of applying AI for a general purpose and hoping for a return on investment, organizations should consider how a domain -specific agents can deal with specific points of pain with measurable effects.

As AI continues to develop from tools we use, for colleagues we work with, differentiation will increasingly be found not in the main models, but in how effective they are trained to understand the specific business contexts and domain knowledge.