Organisations have spent decades building communications infrastructure that works for humans. As AI becomes central to customer engagement strategy, that architecture is becoming a liability.
Every CIO and CXO walking into a board meeting right now is carrying the same mandate: transform customer engagement with AI and do it fast.
Most are also sitting on an architecture that makes that transformation structurally difficult. Not necessarily because of bad decisions, but because of decades of accumulated ones.
Alpana Bajaj, Partner Group Product Manager for Dynamics 365 Contact Centre at Microsoft, has spent 13 years at the company and led multiple products from the ground up. She now leads the growth charter for Microsoft's contact centre business. The pattern she sees across organisations today is consistent.
"Most leaders are realising that accidentally they have built an architecture that is optimised for humans stitching pieces together, as opposed to AI," she says. “Fragmented systems are creating fragmented intelligence."
“Fragmented systems are creating fragmented intelligence."
The Stack Wasn't Built for AI
The fragmentation didn't happen overnight. Telephony came from one vendor, CCaaS from another, CRM from a third. Collaboration tools arrived separately. AI bots were added most recently, layered on top of everything else. Each decision made sense at the time. The cumulative result is a stack where customer context, interaction data, workflow history, and collaboration signals all live in different places.
Bajaj describes a conversation with a customer that illustrates the problem. The organisation had AI running across five different systems in its contact centre ecosystem. When they asked each system a question about the same customer, none of the answers agreed. The data was too fragmented for any of them to produce a coherent picture. Adding more AI tools to a fragmented foundation doesn't fix that. It makes it worse.
Customers Experience Your Architecture Whether You Like It or Not
The consequences show up the moment a customer interaction gets complicated. A customer calls with a complex issue. The AI handling the initial interaction has no context from previous engagements. The agent who picks up has no workflow history. The product specialist in Teams who could resolve the issue in minutes isn't connected to the conversation at all.
"What should have been a two-minute resolution turns into a 20-minute apology tour," Bajaj says. "Customers may not know your architecture, but they absolutely can experience it."
There are visible costs: licensing, implementation, maintenance. The harder cost to quantify is innovation drag. Every new AI workflow, every process automation, every routing improvement means coordinating across multiple vendors and integration layers. As Bajaj puts it: "That's not innovation. That's archaeology." For leaders with board-level mandates to move at the pace AI is evolving, that drag is the real problem.
Why UCaaS and CCaaS Can No Longer Stay Separate
For most of the past two decades, unified communications and contact centre technology lived in different worlds. UCaaS governed employee communications. CCaaS governed customer service. Different teams, different budgets, different vendors.
That separation is breaking down. A complex customer issue today routinely draws in a contact centre agent, a product specialist, a field technician, and a finance expert, all collaborating in real time. The tools those people use to communicate cannot be siloed from the tools used to serve the customer. When they are, the customer pays the price.
AI is accelerating this convergence because it works best when communication and collaboration are connected. The contact centre of the future isn't a department. It is the entire enterprise, connected in real time.
Where to Start
The architecture decisions organisations make now will define how quickly they can innovate over the next decade. In our upcoming video, Bajaj explores what that looks like in practice, including why you can't bolt AI onto a fragmented stack and what connected infrastructure actually delivers. Microsoft's Dynamics 365 Contact Centre, built on the same platform as Teams Phone and Microsoft 365, is designed to be that foundation.
