zeeshan khan
One term is actually changing the way we think about labor in the fast-paced field of intelligent machines, where buzzwords change more quickly than Silicon Valley startups: agentic AI. While the majority of businesses are still learning how to use rudimentary generative AI tools, Zoom is already ahead of the curve, discreetly developing AI systems that proactively anticipate demands, solve complex problems, and act on their own initiative in addition to just responding to commands.
Speaking with Zoom’s Chief Technology Officer, XD Huang, recently, I learned how the firm that made headlines for keeping us linked during the pandemic is now leading a workplace revolution that few anticipated.
Why Is AI Really “Agentic”?
Huang claims that agentic AI is the realization of a goal that has been sought after for generations by tech visionaries. “Software is currently growing up, and intelligence has been an eternal dream for many future generations,” he says. “Ultimately, with the development of generative artificial intelligence large models of languages, the capability achieved the stage that represents the starting point of the application of agentic AI.”
However, what sets really agentic systems apart from the crowded field of artificial intelligence? Huang identifies four key traits: the capacity for independent and proactive activity, memory for a long time, reasoning ability, and task orchestration skills.
See More: Zoom In-Out feature introduced for Videos by Google Photos
“The capacity to independently identify the chance and the setting and the task to help you successfully execute, or obviously, without your authorization, will be fundamentally very important,” Huang informs me. “This is a difficult task that typically calls for several steps and agents. To finish jobs that take more than a few minutes, we must be able to coordinate all of them.”
From AI Productivity Hub To Video Conferencing
Although many people still consider Zoom to be primarily a video conference platform, the business has been going through a major transition. “Zoom is really transitioning from a video conference business that is widely recognized in all nations to an AI-first efficiency company,” Huang says.
Zoom’s AI Companion is at the heart of this change, and Huang refers to it as your “chief of staff.” Huang concisely summed up Zoom’s new ambition with its AI assistant: “We are moving from meeting to milestone. We are holding these sessions in order to reach our next goal.
This progression is quite logical. In the end, meetings are about achieving goals, not just having a talk. Zoom is reorienting its entire product philosophy around results rather than activities by concentrating on this essential principle.
AI Companion: An Overview of the Workplace of the Future
What are the current capabilities of Zoom’s AI Companion? Much more than you may think. It does more than simply record and transcribe conversations during meetings; it also actively listens, makes context-based decisions, and assigns follow-up activities. Huang explains how it designates responsibilities to particular team members and, more and more, is able to carry them out on its own.
Huang gave a personal story that was quite illuminating: “I got a piece of email. I have three people I need to meet with, but I don’t have a slot. I can just request that AI Companion set up a meeting for my available time slot. AI Companion will take care of that for me.
The fact that Zoom’s artificial intelligence (AI) features now go beyond virtual meetings is even more astounding. “I went to a meeting of the customer advisory board. I just placed my iPhone on the table, switched on the AI Companion, and after the meeting—which had around 70 people in the room—I received the minutes just like I would have from a virtual meeting.”
Most unexpectedly, Huang disclosed how he created his own evaluation of performance using AI Companion. “All of the docs I produced are accessible to AI Companion. I would say that AI Companion completed 80% of the assignment, which includes all of my knowledge work, decisions, and recommendations.
Small Language Models’ Power
By creating specialised small language models (SLMs), Zoom has adopted an intriguing counter-approach to the industry’s preoccupation with ever-larger language models. In particular, they have developed a model with two billion parameters that is remarkably efficient.
Huang uses a fitting example to explain the approach: “I would say we might have an encyclopedia that covers everything, but we also have specializations.” We delve deeply. Both are necessary for society. This strategy exemplifies what Huang refers to as “T-shaped knowledge”—deep vertically experience combined with extensive horizontally awareness.
The outcomes have been outstanding. “The level of quality in this 2 billion small language model, based on the overall standard, is as good as the bigger language model which existed two years ago,” Huang says. These tiny models can actually beat much larger competitors at a tenth of the cost when tailored for particular tasks, such as translation.
“I think that everyone will likely have a lot of agents in the future perhaps ten or twenty. A little language model tailored to each agent can be installed, according to Huang’s prediction. “And the bigger model of language will bring those small models of language together in order to orchestrate them seamlessly.”
The Opportunities For Collaboration In The Workplace
Huang goes back to his main point when asked how agentic AI will change collaboration at work in the upcoming years: “From meeting to milestone.” You can be more driven and reach your milestone much sooner thanks to productivity.
Read More: Leveraging agentic AI to power the next evolution of Zoom
He depicts AI assistants that, while maintaining human control, keep an eye on pertinent data around-the-clock, carry out independent study, and then come back with suggestions. “Consider that a hundred individuals are working for you. Following this meeting, they will examine the topics we discussed, conduct thorough study, gather data, and provide the best advice they can.
Conclusion:
The world of work is evolving faster than ever, and Zoom is clearly not sitting still. By leaning into the power of agentic AI, it’s moving beyond being just a video call platform—it’s becoming an intelligent partner that helps teams actually get things done. From organizing your calendar to pulling insights from meetings you barely had time to attend, Zoom’s AI Companion is turning small daily tasks into automated wins. And maybe that’s the real magic—making work feel less like work, and more like progress.
FAQs:
Q1: What makes Zoom’s Agentic AI different from other AI tools?
The majority of AI tools react when prompted, whereas Zoom’s Agentic AI acts on its own initiative. In addition to making decisions, it listens, comprehends context, and even takes action on your behalf by scheduling meetings, compiling notes, and delegating responsibilities.
Q2: Can I use Zoom’s AI Companion outside of video meetings?
Yes! Zoom’s AI Companion now works in physical meetings too. Just set your phone on the table, hit record, and the AI will deliver a full summary and action items afterward—no typing or note-taking needed.
Q3: Will agentic AI replace human workers?
Absolutely not. Enhancing your abilities rather than replacing you is the main goal. Consider it a very intelligent helper that relieves you of tedious chores so you can concentrate on planning, coming up with new ideas, and making decisions.
