No Jitter asked Brent Kelly of Omdia for his thoughts on the year that was and the year that will be with respect to the ongoing merger of unified communications as a service (UCaaS) and contact center as a service (CCaaS) capabilities into a single platform and the integration of generative AI (Gen AI) capabilities into these same platforms.
No Jitter (NJ): Last year we talked about the merger of UCaaS and CCaaS capabilities. That trend has continued during 2024. What are your thoughts on why that’s happening?
Brent Kelly (BK): Well, quite frankly, the UCaaS vendors have reached the end of what they could offer, with respect to telephony, meetings, messaging and so forth. And they're all pretty much the same. If I use Cisco, Microsoft, Zoom or RingCentral – or any UCaaS product – the capabilities are very much the same. They're all good at doing voice messaging and meetings, for example.
So the vendors are looking for some other way to generate revenue. It's natural to look for an adjacency. And because they have a communications platform already, one of the most natural adjacencies is contact center. The UCaaS vendors who have what I’ll call, for sake of discussion, a ‘telephony-enabled UCaaS’ (because some of the vendors have a UCaaS play that is not telephony enabled) can leverage those capabilities to help them differentiate and try to increase their revenues and viability in the market.
There's something else that's at play here, as well. Prior to the existence of CCaaS – that is, ‘as a service’ – companies would buy a PBX, and then they would buy a contact center. And 60 to 70% of the time, companies would buy their contact center from their PBX manufacturer. That worked well in the days of on prem because the contact center played well with the PBX, it was easier to deploy and since both were from one vendor, the company had the proverbial ‘one, one throat to choke.’ That mentality still exists.
We did a survey in 2023 where we asked people who already had a UCaaS platform if they could buy a contact center from the same company as provides their UCaaS, would they do so? What was stunning was about 70% of the people said, ‘yes, we would.’ So, there's this long-term dynamic here that people would like to buy, let's say their communications from a single provider which is still valid in this era of cloud-based communications.
NJ: What’s your take on some of the generative AI capabilities that are being added to UCaaS and CCaaS platforms? Is it along the lines of what you just described – vendors seeking differentiation and revenue growth?
BK: As I mentioned, the UCaaS companies need a way to continue to grow, but they've also recognized that that they need other capabilities. I started covering this market before it was called UCaaS – before it was called unified communications, even.
Effectively what we had were companies that had a PBX and then realized, oh, maybe we should add in some messaging. That's kind of how Cisco got into the market. Cisco had a PBX, then they bought Webex and got into meetings and then video. Microsoft first entered the market with a mostly messaging product and soon realized they needed to compete with what was happening with Webex and others, so they bought a meetings company and added the ability to do online meetings. Then they added voice and then some contact center capabilities. RingCentral started with phones, then moved into messaging and are now doing video and events. Zoom started with video only, then added meetings, and so on.
So, I think this is just a natural progression. As companies develop these products, they find adjacencies where they can also play. One of the other adjacencies in the UCaaS area, aside from contact center, is this notion of events and webinars, which most of the companies have expanded into providing.
With respect to generative AI and summarization and those kinds of features, and how they’re being infused into a lot of applications, my sense is that knowledge workers will mostly use meeting summarization. The creation capabilities – Copilot for Word, Teams, Excel – well, I've had access to them all for over a year now, and Copilot for Excel or PowerPoint just don't add value for me. But Copilot for Word, on occasion, adds a lot of value for me.
In the contact center, call summarization and agent assist will be the most-used capabilities. Even though you hear people saying that they’re going to omnichannel and voice calls are decreasing, the fact is, voice calls are still 60 to 70% of the interactions with the contact center, and so they need this.
Also note that vendors will charge for any AI in the contact center, but on the UCaaS side they give summarization away for free – at least RingCentral, Zoom and Cisco Webex give it away free. Microsoft doesn't, but all these companies will charge for AI summarization, and agent assist, on the contact center side.
But I think there's this idea percolating up that large language models – from OpenAI, Anthropic and others – are overkill for most of the things that people need. What they really need is a small language model that has been trained on what’s important for what they're trying to do in their business. So I think the use of small language models that you can use retrieval, augmented generation [RAG] on and ‘infuse’ into them with your own data and information. I think those models will be far cheaper and far more accessible.
Companies will be able to develop the smaller models on their own or have somebody else develop them for them, so that they're not subject to the charges of these large [LLM providers], plus they can keep all their own data. I know that there's security guardrails in place with all the [LLM and cloud] providers but there's still something about keeping your own data. So, the impression I have is that companies will get AI more into their own hands, as opposed to relying on the vendors to provide it all for them.
NJ: Crystal ball time: Any trends you see emerging in 2025 related to UCaaS
BK: I keep saying that there's going to be price compression, and I've said it for years, and it really hasn't happened. It really should happen, because as these companies get more and more customers, their cost of servicing an individual customer goes down because they're on shared infrastructure and so forth. But they keep adding capabilities to their different SKUs so that they can keep those prices high. I'll predict it again that ultimately, whether it's this year or in a couple years, we'll see some pricing pressure, because the greenfield opportunities are diminishing and shrinking with respect to UCaaS.
As competition heats up, and particularly in the contact center as we get new players that are now focusing on contact center. For example, Dialpad now says it’s a contact center company. Nextiva does, as well. And people are saying, well, contact centers where the market, where the money is, and to some extent that's true, but, but that the availability of greenfield opportunities, even there, will shrink. And so ultimately, price compression is one.
I think there may be some consolidation, as well. I estimate that a vendor will need 8 to 10 million subscribers, as a minimum, to end up being a long-term player in the UCaaS market. There's not a lot of companies that have that yet, but a few are getting close.
