Enterprise Connect 2026 brings AI from hype to reality

IT leaders are looking to turn the AI hype into reality for their unified communications and customer experience systems by seeking practical advice on how to bring AI capabilities to employee and customer workflows while delivering measurable outcomes, such as increased productivity and reduced costs. UC and CX vendors, too, are facing pressure to demonstrate their relevance in an AI-driven world, with renewed focus on vertical offerings and differentiating from AI vendors like Anthropic and OpenAI.

At Enterprise Connect 2026, which takes place March 10-12 in Las Vegas, conference attendees will explore the practicalities of adopting AI, including developing AI governance and security policies, wrangling AI agents and combating bias.

They will also discuss continuing strategies to support hybrid work, including meeting room design and infrastructure updates. CX leaders, too, will gain insights into enhancing the agent experience and modernizing the contact center in the age of AI.

In this Q&A, Metrigy analyst Irwin Lazar discusses the topics that will be top of mind for IT leaders navigating the changing UC and CX landscapes.

Editor's note: The following interview was edited for length and clarity.

What are the big themes at Enterprise Connect this year?

Irwin Lazar: The first and maybe the most important one is the impact of AI on the vendors at Enterprise Connect themselves. There's been a lot of chatter that Claude code and AI will essentially make SaaS obsolete. I think a lot of the vendors are reacting to that. I've spent the last couple of weeks with Zoho and RingCentral, and that was the first thing that they wanted to address -- that SaaS companies can maintain relevance and how they can differentiate. That's going to be the talking point that every vendor there is going to have.

Irwin Lazar

We've spent the last couple of years talking about AI and the hype of AI. I think we're moving more into the practical benefits and solutions. What I'm hoping to see is companies coming out with specific AI features that allow people to customize workflows, that allow them to add specific capabilities into their products that are backed by quantifiable KPIs, that can increase productivity, reduce operating costs and improve revenue.

The third area is around vertical solutions. The UCaaS space has become largely commoditized, so how do vendors come out with specific solutions that benefit specific industries -- retail, financial services, hospitality, healthcare? I think you'll see a lot more focus on those kinds of solutions.

AI governance has been a big topic. Is that something IT leaders still struggle with?

Lazar: There are a number of different areas we see companies really struggling with, and it's been a hindrance to the successful implementation of AI. The first is data classification and data governance, so ensuring that if I put my data into an AI LLM [large language model], somebody querying that data can only see what they should be allowed to see.

We published a study last fall on AI governance and security, and we found that data classification and data leakage were the biggest factors inhibiting deployments of AI -- along with concerns about the accuracy of data. If you're relying on data scraped from Reddit, X and other public services to answer customer inquiries, you may get some oddball responses.

Another issue that's starting to increase is the governance of AI agents. As companies are starting to roll out agents, you could end up with hundreds of thousands or tens of thousands of agents. Understanding what's out there, how it's being used, whether people are duplicating efforts -- how do you catalog and make the work that others have done available to the rest of the company?  

I'm doing a session on Wednesday morning with Martha Buyer, who's a telecommunications attorney, focused on AI bias. AI models can be biased based on a variety of issues. They could be biased based on how they were trained, whether socioeconomic concerns influenced how they were constructed, or based on how they're cataloging decisions that were made over 20 years ago, when society may have been different than it is now. How do you make sure those AI models are removing bias in their responses? It's an interesting topic, and it's the first time we've done anything like that at Enterprise Connect, so I'm excited about that session.

Deepfakes also seem to be a hot security topic this year, too.

Lazar: We're doing a half-day workshop focused on AV [audiovisual], and I'm moderating a discussion around security, and that's a huge issue. We see a lot of concern in the contact center. Companies are spending an awful lot of time trying to validate customer identity and protect themselves against those deepfake attacks, where if I've got your voice or a snippet of your video, I can create a fairly sophisticated attack. Then ensuring that people aren't essentially creating deepfakes that are impersonation attacks meant to embarrass companies.

Identity protection is a huge issue. We see that companies are actively trying to look at internal help desks. If somebody's calling in to tech support, and they have obtained a stolen credential or cookie, they're trying to convince the tech support person to reset a password and use a social engineering attack by impersonating somebody's voice.

These attacks get really sophisticated. The other thing we'll talk about is the flip side of using AI to protect yourself. AI is pretty good at detecting these kinds of attacks and potentially automatically mitigating against them.

What questions do IT leaders want answered at the conference?

Lazar: AI overload is the challenge that we see companies face, and there are so many different AI capabilities coming out from so many different vendors. How do you manage it all? How do you eliminate redundancy? What's the single source of truth for data?

If I'm in a multi-vendor shop, they all have their own AIs. Is there a way to link them together? We're seeing a lot of interest in MCP [Model Context Protocol] servers and ways to allow AI models to share data with one another. And then I think, you know, what becomes what is the right approach.

We see companies saying, "I'm going to go with Anthropic or OpenAI as my primary AI platform, and I want to pull all my data from all my different applications into that." That takes that data out of the vendors that really want to deliver AI solutions of their own and eliminates their ability to differentiate. That's something companies are going to be asking about: How do I get out of these silos of AI, and what's the practical ROI? We've heard for the last two to three years now that AI is everywhere. It's all anyone talks about, but how do I translate that into a practical business benefit?

I think 2026 is the year of rollout versus 2025, which was the year of kicking the tires and trying to figure out what direction we want to take. I talked about security, compliance, governance being a big issue, but I think just that overall ROI story, what platforms make sense, what models make sense -- I think that's where companies are struggling right now.

You're also speaking on the 911 panel again this year.

Lazar: This has got to be the 9th or 10th year we're doing a 911 panel -- it's still an issue. Every year we survey our research participants and ask where they are related to compliance of Kari's Law and RAY BAUM's Act. With Kari's Law requiring essentially direct dial to 911, and on-site notification with RAY BAUM's Act governing location information that you share. We find that 20-30% of companies will tell us, "No, we don't think we need to be compliant; we're grandfathered in with older systems." We want to address that.

We also want to talk about Next Generation 911, which is becoming more and more of a reality. You're seeing efforts underway to retire some of the legacy systems to support [Session Initiation Protocol] SIP-based communications with 911 centers. It's one of those sessions where every year we go, do we really have anything to talk about? And then we end up with a full room and more questions than we have time to answer.

Katherine Finnell is senior site editor for TechTarget's unified communications site. She writes and edits articles on a variety of business communications technology topics, including unified communications as a service, video conferencing and collaboration.