2025 in Review: The Year CX Got Interesting Again?

2025 was the year everyone claimed to be AI-first. It was also the year many organisations discovered that saying it and operating it are two very different things!

But beneath the hype, something real happened: CX actually moved forward. Meaningfully. Practically. And at speed, for once! It stopped being treated as a collection of tools and started being treated as a system: Connected, intelligent, managed, and increasingly inseparable from the rest of the business.

From our perspective at Acceleraate, and echoed by Zoom, this was the year CX got interesting again! Here's our roundup of 2025, and our take on what it means for 2026.

Unified CX stopped being optional

For a long time, CX investment followed a familiar process: Add a chatbot. Patch in analytics. Bolt on an integration when something breaks. Over time, organisations ended up with fragmented stacks that technically worked, but drained performance through friction, duplicated effort, and inconsistent experiences. But in 2025, that approach started to unravel.

As Ben Neo, Customer Experience Sales Leader EMEA at Zoom, observed:

2025 was the start of customers realising the power of unified CX platforms to seamlessly connect the entire customer journey across their organisation, rather than continuing to invest in point solutions that unintentionally introduce friction and fragmented engagement.

This changed the conversation. Buyers stopped asking which feature was best and started asking how everything worked together. How context moves across channels. How AI works across the journey, not just at the edges. And how the contact centre connects to the rest of the organisation.

Unified CX is no longer something to aspire to in the future. It's become the baseline expectation.

AI stopped being Sci-fi

If 2024 was the year AI amazed people, 2025 was the year it was put to work. And the real impact came from how AI was embedded, governed, and connected to real workflows, and whether it reduced effort rather than introducing more complexity.

As Ben Neo put it so succinctly:

Satisfaction is retrospective. Effort is predictive.

Across the UK and EMEA, the most mature organisations stopped focusing purely on how customers felt at the end of an interaction and started examining how hard that interaction had been in the first place:

• How many steps were required.

• How many handoffs occurred.

• How often customers or agents had to repeat themselves.

With AI embedded across routing, self-service, agent assistance, analytics, and orchestration leaders finally began to see the hidden effort inside their operations. More importantly, they started designing journeys that removed that friction before it turned into dissatisfaction or churn.

What changed behind the scenes was just as important

This shift wasn’t only visible at the experience layer. It happened deep in the engineering foundations too. According to Ben Jemison, CTO at Acceleraate, 2025 fundamentally changed how integration engineering itself is done:

Using AI as part of our software engineering process for IntegrationHub - in a controlled and secure way - helped us cut through a lot of the heavy lifting that often slows teams down. It’s helped us interpret complex data, spot relationships across events, and surface what actually matters, instead of drowning in noise.

The significance here isn’t that AI started writing code. It’s that it removed friction from the engineering process itself. It reduced manual effort, clarified complexity, and freed-up our teams to focus on what really matters: architecture, security, resilience, and outcomes.

As Ben put it:

It’s made our platform smarter without making it more complicated; delivering cleaner, faster, and more reliable integrations for customers.

In a year where CX platforms were consolidating rapidly, integration quality became the difference between acceleration and abject chaos.

Hybrid work stopped being a compromise

Another long-running debate quietly disappeared in 2025: Human versus bot.

The better question, as Ben Neo framed it, became: What is the best possible resource for this moment?

Hybrid work matured. Virtual agents became first-class participants in the workforce. AI began to dynamically decide when automation was appropriate and when empathy, judgement, or creativity delivered greater value. Crucially, bots stopped being treated as free capacity and started being governed, measured, and optimised like any other part of the workforce. The organisations that are pulling ahead have started managing humans and AI together as a single, adaptive workforce.

Regulation forced better design

2025 also brought a dose of reality. The EU AI Act, evolving UK regulation, data sovereignty requirements, and increased scrutiny around explainability and governance made it clear that AI in CX could no longer be experimental. For some organisations, this was uncomfortable, but for others, it was validating.

Platforms and partners that had prioritised unified architecture, clean data flows, and embedded governance found themselves able to move faster. Zoom’s federated, AI-first approach, combined with Acceleraate’s compliance-first delivery, meant customers could innovate with confidence rather than caution.

Danielle McLoughlin, General Counsel at Acceleraate, explains:

2025 was the year wider teams had to engage with the EU AI Act, or at least the high-level overview. For many organisations, it turned out to be less dramatic than anticipated. For those planning ahead, building for agility, transparency, and ethical AI use is good preparation, and a customer expectation.

Looking to 2026, we expect two shifts: customers will increasingly want to understand how AI-decisions are made, not just what is being delivered. And organisations investing in adaptable, ethics-first frameworks will be far better positioned as enforcement and regulatory guidance evolves across jurisdictions.

From CX as a function to CX as infrastructure

As Ben Neo observed:

The real AI revolution in CX won’t be technological - it will be organisational.

In 2025, the most advanced organisations broke down the walls between insight and action. CX data was no longer being reported after the fact and started being used as live operational insights. Frontline interactions informed decisions across operations, product, and leadership. And at that point, CX stopped being a department and became fundamental infrastructure.

Looking ahead to 2026: From AI-first to AI-native

No annual CX review would be complete without a few predictions, and 2026 is shaping up to be a year where several long-promised shifts will finally become unavoidable.

The biggest change won’t be more AI in CX. It will be AI everywhere, by default.

In 2026, we expect organisations to move decisively from AI-first to AI-native. Conversations won’t just be captured; they’ll be enriched, analysed, summarised, and actioned in real time, without needing to be switched on or “added” to a process. AI will simply be part of how work happens, across CX, operations, sales, product, and service delivery. And with the recent launch of AI Companion 3.0 Zoom is extremely well positioned to support organisations through this evolution.

And that’s a critical point: AI will stop being confined to the contact centre. Customer interactions will increasingly feed wider organisational intelligence, informing staffing decisions, product improvements, service design, and commercial strategy. CX won’t just consume AI; it will power the rest of the business with insight.

This will accelerate another overdue breakup: The slow, inevitable end of siloed and legacy CX technology. In 2026, the tolerance for fragmented stacks (on-prem telephony, bolt-on CC platforms, and brittle integrations) will be gone. Not because IT suddenly became radical, but because unified, cloud-native platforms have proven they are faster to deploy, easier to govern, and cheaper to evolve. As we like to joke internally, no CIO is approving another on-prem telephony refresh if they don’t absolutely have to!

As these platforms unify, how success is measured will change too. Traditional contact centre metrics like AHT and occupancy won’t disappear, but they’ll stop being the headlines. In their place, organisations will focus on measures that reflect modern CX reality: AI quality, orchestration effectiveness, agent enablement, and - as Ben Neo highlighted - effort reduction. Satisfaction tells you what happened yesterday. Effort tells you what will happen next.

In 2026, the real competitive advantage will come from how well humans, automation, and data are managed together. AI will dynamically decide when to automate, when to escalate, and when human empathy delivers more value. Not just emotionally, but economically. Bots will be governed and optimised like a workforce, not treated as free capacity.

Finally, 2026 looks like a breakout year for UK and EMEA public sector AI adoption. With governance frameworks maturing, regulatory clarity improving, and proven deployment models emerging, the conversation will move from “can we?” to “how fast can we do this responsibly?”. The floodgates won’t open recklessly, but they will open.

CX in 2026 won’t be a function, a toolset, or a reporting layer. It will be an AI-native, integrated operating system for the organisation. And CX leaders will sit closer to the boardroom than ever before.