From AI agents to cloud telephony, Zoom executives discuss company's next act

Indian Express

18-06-2025

Zoom wants to be much more than a video chat platform. It is taking on Google and Microsoft in enterprise software by shipping productivity tools such as Zoom Documents, Zoom Team Chat, Zoom Whiteboard, etc. It has also rolled out Zoom AI Companion and is integrating AI into all levels of its product suite.
Zoom is also looking to strengthen its foothold in India by expanding its cloud-based enterprise telephony service called Zoom Phone to four additional licenced service areas across the country. The company further launched Contact Center here that enables customers to reach out to Zoom's enterprise clients through a wide range of channels, including voice, video, virtual agents, social media, email, and messaging apps.
Security is central to everything Zoom does, Steve Rafferty, head of APAC and EMEA at Zoom, told indianexpress.com in an exclusive interaction in New Delhi last week. Echoing this, Sameer Raje, the general manager and head of India & SAARC region at Zoom, highlighted the company's federated approach to AI that empowers its customers with greater control, privacy, and choice.
Here are the edited excerpts from the interview:
Zoom became a household name during the COVID-19 pandemic. Since then, it has pivoted to becoming more than a video conferencing platform. How does Zoom Phone and launching Zoom Contact Center in India further that vision?
Steve: If you think of Zoom as a business communications platform, then video and meetings was the first element of that. Lots of organisations enter from different pivots. So we were meetings first, and now we've got chat, Contact Center, Phone, Clips, Documents, and so much more. We're trying to give our customers and partners one place to meet effectively, whether that's message, video, or phone. Whether it's internal communication or messaging externally with customers or suppliers, Zoom will give you that unified platform experience. That was always the plan. We just started with meetings.
You mentioned that getting Zoom Phone regulatory compliant in India was an elaborate process. What were some of the terms and conditions involved in the DoT authorisation? Also, what do you think needs to change in terms of the eligibility conditions for cloud-based telephony services?
Sameer: Authorisation is basically the license. It's the UL access BMO license which we took in all the six telecom circles [Mumbai, Bengaluru, Chennai, Hyderabad, Andhra Pradesh, and Delhi-NCR]. And each of the circles have their own requirements as well as technical call flow, approvals, so on and so forth.
It's not as simple as building it once for one telecom circle and having it work everywhere. Logically, yes, the product might be the same but each circle requires its own set of regulatory approvals. When you have a global product, the DoT requirements are very stringent. For instance, all equipment must be certified by DoT. There are specific certification programmes and you cannot just use any hardware. From the hardware to the call flows, we had to change some of the elements of the product to make it compliant within the country. Hence, it was time consuming.
Not only that, each telecom circle needs to be connected not only to the Zoom cloud but also to local telcos across the region and country. That includes setting up incoming and outgoing interconnections, which is a complex and time-consuming process.
But that doesn't mean that you cannot use Zoom Phone anywhere outside these circles. Right now, these six circles are where you can procure the license from, but you can travel anywhere and accept calls and make calls to anywhere in the country.
It's been more than a year since Zoom Workplace was introduced. How's the response been from enterprise clients compared to everyday users in terms of adoption and overall experience? How much of that has translated to sales?
Steve: It's been very strong across all of our regions, especially internationally.
What we're seeing now is different entry points for customers. Some customers may come in to start from a Contact Center journey because they've got a particular customer requirement, and we fix that. Then they'll adopt Zoom Phone or messaging.
The workforce is changing at the moment. We're seeing a lot younger, more diverse workforce come in, and they've really taken to the Zoom platform because you can communicate by chat, you can pick up the phone, you can send a clip, you can join a work meeting, engage with customers, and so on.
There's so many different angles to procure information. We've got control of so much data. A business can take that data and deliver better experiences to their customers.
Have you noticed differences in how AI features are being adopted across specific use cases like customer support, live translation, meeting summaries, etc?
Steve: There's two sides to that. There's the internal day-to-day job, and there's how you communicate with customers. We were with a customer last week at a sports shop. What they want is a specific chatbot for their clients, and they also want to be able to escalate that call really quickly to a human depending on the requirements. So if somebody wants to procure a tennis shoe, it is dead easy to handle. But if somebody wants bespoke running shoes, that's a specialist treatment. That's what we're seeing from a customer perspective. Zoom Workplace allows you to move those conversations around.
Sameer: I'll add one critical point to that. We include Zoom Workplace in our license at no extra cost. Any other platform or AI engine may charge $20-$30 per user and the CEO or senior leadership may go for it thinking its the in thing. But are they really going to be that productive? Or does it make more sense to spend that $30 at the lower level so that the support agent can be more productive using AI on Zoom? They should be utilising AI where it makes better sense. It could be customer support or sales. Zoom Workplace can be rolled out across the organisation. One of the other areas that has been very positive to me is having all of that in one location. All the integrations on Zoom Workplace makes life so much easier and gives us hours back every week.
While LLMs (large language models) can be quite convincing, trust is still a barrier to enterprise adoption. How do you solve for that?
Steve: Security has to be at the heart of everything you do, whether it's AI or whether it's a basic implementation of any technology. We work with our customers, from CISO all the way down, to make sure that the data they're exposing is the correct data. We work with a security bank because you've seen from some of our competitors, how they've exposed people's salaries all the way through a business. We start with the security first? What's the outcome you're looking for, and how can we give you a secure outcome? And then the platform gets deployed based on that.
What do you think about the role of AI agents on your platform, especially in automating enterprise workflows? Do you see them evolving into a meaningful revenue stream?
Steve: It's about complementing the human workforce. It's not about replacing them. You've seen a lot of press with some organisations saying 'we're going to slim down our workforce. We're going to stick AI agents out there', and then it's difficult. What we do is, is, we speed up that journey for the customer. If you think about booking flights on air travel, for instance, we can pretty much run that whole system through a virtual agent.
Sameer: If you look at our AI Companion, it's been embedded in a couple of products that we've introduced like Zoom Whiteboard where you can draw a circle and the AI will actually make it into a perfect circle. Zoom Workflows is also in beta right now. Beyond that, we also introduced Custom AI. Let's say you're in a pharmaceutical industry and you're using a particular terminology. You want the AI model to use that terminology, right? So it allows you to create your own language within that.
What is Zoom's approach to AI development?
Sameer: We have a federated model. This means we use our own LLMs as well as other industry-standard LLMs (OpenAI and Anthropic's models). That's how we shift between low-cost and high-cost GPUs. We also permit customers to bring their elements to a certain extent. That's a chargeable service. We are helping organisations break down their silos and improvise employee experience.
We're quite happy for the customer to work where they want, and we support them with all of our tools and our AI and and our platform underneath. It's quite a different model to others.
We've also made a commitment that we won't use customer data to train our models.
But you're also in competition with these big tech companies, right? Most of their AI services seem to be bundled into their office suites.
Steve: Yes but we don't walk into a Google or a Microsoft customer and say, 'What a terrible decision. You need to rip it all out. Stick Zoom in.' We go: 'Congratulations on your decision. Now, let's make sure you get the best value for money'.
Not everything that's bundled works, and you're seeing how those bundles consistently change. What we allow you to do is to take the best video platform and plug it in, as well as the best phone platform if you want to use it. We've got Contact Center that you can plug in to make your Google or Microsoft Office environment sing much better.