We are experiencing a generational transformation of service delivery, and it’s not just about artificial intelligence (AI). I believe we are also seeing a reordering of how software providers organize the tools used in contact centers. We call it Contact Center as a Service (CCaaS) because that’s what people are used to, even though “as a Service” doesn’t accurately describe an environment where around half of seats are still on-premises. Whatever it’s named, the model is collapsing and I find it unlikely that the providers and products of the 2030s will look much like today’s default.
I want to highlight three ideas that suggest the kinds of changes we are in store for.
First, the shift away from the ACD as the foundation of the contact center stack. Traditional ACDs were magnificently effective at creating speed, scale and efficiency around voice routing. But they were also hideously expensive; closed and proprietary; and very voice-centric. The move to the cloud opened some of that up, making multichannel routing more accessible and introducing a much faster cadence of feature development. That was all well and good, but it also sometimes made it harder to connect the CCaaS to enterprise and back-office systems and to achieve the kinds of deep customizations and control that were standard in on-premises tools. It also introduced questions that went beyond the CCaaS itself, like where does my data live and how is it controlled?
I think we’re moving towards a future state where CCaaS becomes more hybridized with enterprise orchestration efforts. For instance, instead of relying solely on the CCaaS provider’s logic, enterprises place a middleware or orchestration fabric in front of CCaaS. This controls management of the overall customer journey, data movement, and channel switching. And the centralization of ownership over customer and interaction data moves outside the center into a coordinated data lake or warehouse managed by data teams, not contact center teams. So, the CCaaS generates calls, chats and transcripts, but the enterprise ingests and analyzes them alongside CRM, ERP, and marketing data.
At the same time, self-service is starting to break apart from traditional call handling due to conversational AI, which is represented by a much more diverse provider landscape. Central control is being sacrificed to attain better AI outcomes. The overall result is more consistent customer experience (CX) orchestration across digital and voice channels, alongside enterprise-owned data and AI resources to ensure compliance and control.
Adding conversational AI increases the orchestration complexity of the interaction environment. In effect, it turns integration into the defining differentiator, encouraging buyers to ask: can vendors provide the unified data models, orchestration and API support needed to transition from the response-driven contact center of the past to the orchestration-driven contact center/CX of the future? It forces providers to focus less on interaction handling and more on experience orchestration.
The second thing to consider (and an outgrowth of the first) is how enterprises structure their contact centers internally. We’re going to start to see the end of the contact center as a defined entity within businesses. Fewer people will be needed to handle digital and AI interactions, while at the same time CX teams outside the contact center will be more involved in orchestrating, personalizing and measuring the outcomes of complex service interactions. In fact, “service” interactions will be more of a mélange of upsells, cross-sells, problem solving and asynchronous ongoing conversations. That de-centralizes the center as the locus of service delivery.
And third, the hyperscalers are taking the basic level of CCaaS away from low- and mid-range providers. Some giants excel at unifying contact centers, AI and marketing/sales applications, and they have extraordinarily deep pockets to outspend legacy providers in R&D and acquisition, especially around AI.
They can supply a majority of the underlying traditional tech stack, much of which has emerged as new in recent years, and they outright control some of the core elements that CCaaS providers have to purchase or access: cloud compute, storage and networking; foundational AI models; and AI-agent infrastructure, among others. So, traditional CCaaS providers risk becoming a thin software layer sitting on top of infrastructure, AI and communications services supplied by a much larger platform company.
Hyperscalers don’t necessarily have to maximize CCaaS software margins. Contact center workloads also generate tons of consumption of the services they excel at. And so, your traditional CCaaS provider won’t be able to sustain premium pricing for things like omnichannel routing, summaries, basic virtual agents, standard analytics, etc. These become commoditized, and legacy CCaaS providers get stuck on a treadmill trying to differentiate based on features of uncertain value and performance.
The transition from seat pricing towards consumption and outcomes also generates pricing pressure. Seat growth weakens as AI handles a greater proportion of interactions, and a provider that is dependent on human-agent licenses can automate away part of its own revenue base. Enterprise customers will start to expect to pay for completed work rather than for dormant capacity. So pricing gradually shifts toward metrics such as automated resolutions, successful journeys, handled interactions or consumed platform resources.
CCaaS providers therefore have to redesign their revenue models before AI-driven agent reduction does it for them.
So, what does “CCaaS” (best described as “all the software used in customer service delivery, regardless of platform”) look like in the 2030s? Besides the hyperscalers, I think you have some (not all) of the current legacy companies still in the mix, differentiating on more sophisticated applications plus the size and complexity of deployments. These providers would have to be additive to the underlying hyperscaler infrastructures, not directly competitive with them. Then you’ll have a set of companies offering to power contact centers using CRM and service workflow platforms, all coming at service delivery through the customer record and case management layer.
And the rest of the market will fall into specialties—particular regions, vertical use cases, and for enterprises that have a baked-in bias towards developing and customizing their own applications atop existing CCaaS or Communications Platform as a Service (CPaaS) platforms.
The middle of the market thins out. The largest independent CCaaS providers become orchestration and experience-management companies, while smaller ones specialize,
ISG Research already asserts that by 2028, providers outside the legacy contact center space will have captured 25% of the market for contact center seats. That’s just the beginning.
Hyperscalers don’t have to replace every CCaaS provider to take much of the industry’s economic value. What they can—and I think will—do is capture the fastest-growing value pools in AI inference, data, cloud infrastructure, security and developer platforms, all while leaving incumbent providers responsible for the narrower, more operationally demanding application layer. The CCaaS companies that survive will be those that own customer service workflows, orchestration expertise, industry context or measurable business outcomes rather than merely the communications stack.
Regards,
Keith Dawson