It’s Time for a 2025 AI Heat Check: Is it Delivering or ‘Overhyped’?

As AI hype reaches fever pitch, enterprise tech leaders face a new kind of pressure, from the boardroom to the front line. But are AI copilots and assistants truly ready to deliver enterprise-scale ROI, or are we still deep in the experimentation phase? We asked leading voices in unified communications and collaboration for their take

6
It's Time for a 2025 AI Heat Check: Is it 'Overhyped' or Delivering?
CollaborationNews Analysis

Published: June 2, 2025

Kieran Devlin

Almost three years into the post-ChatGPT AI explosion, the technology continues to be the defining trend in not only communications and collaboration but also the business world at large.

You can’t move an inch for big and bold promises right now, from copilots to take the admin off your plate to assistants to make meetings smarter, bots to boost productivity, and AI platforms to elevate the fundamentals of workflows. Yet for all the noise, quieter, more urgent questions are emerging in the minds of IT leaders and tech execs: What’s actually real, and where’s the tangible value?

Behind the dazzling demos and slick marketing, many teams are still labouring in pilot purgatory. Shadow AI is creeping in under the radar, and diligent governance far too frequently feels like an afterthought. While vendors race to ship AI features at speed, CIOs are being asked to justify spending, mitigate risk, and somehow transform potential into real business value, without the luxury of patience or longer-term strategic plans.

So, where are we in the AI hype cycle? Have we hit peak expectation? Or are we finally entering the phase where AI delivers something concrete?

To find out, UC Today canvassed its panel of analysts and experts from across the unified communications and collaboration space to take stock, frankly and pragmatically, of where things stand. As major AI announcements from all the big players in recent weeks indicate, including Microsoft at its Build showcase and a seismic Zoom partnership revealed at Google I/O, the hype isn’t going away anytime soon. But the more compelling story is what happens next.

Where AI Actually Stands for Organisations, From Pilot Fatigue to Production Pressure

Blair Pleasant, President and Analyst at COMMFusion, recently wrote a report breaking down exactly this topic and is understandably conflicted in her assessment: “Let me start by saying: I’m very pro-AI. I’ve seen some amazing use cases, especially in healthcare, that show how transformative it can be. But I’m also concerned that it’s become overhyped.”

Companies are rushing to deploy it without proper strategies, and I think we’ll see some poor and even dangerous implementations in the coming months and years.”

Pleasant suggested that the industry itself is complicit, with the numbers posted by vendors and analysts often inflated wildly. She highlights claims like “98 percent of contact centres are using AI” or “90 percent of small businesses are using it daily,” but most small businesses she talks to don’t even understand what AI is, let alone use it meaningfully. One worrying outcome from this discourse might be that these inflated stats pressure companies into adopting AI prematurely.

“At customer conferences, I always speak to the users, not just the ones on stage,” Pleasant added. “Most are still in the early stages: doing POCs and rolling it out to a department or two. Very few have fully deployed AI solutions.”

“I also surveyed consultants and resellers recently. One said that when vendors speak off the record, they admit AI adoption is much slower than hoped. Many clients don’t know how to begin or are afraid of getting it wrong. Another noted that large enterprises are further along, while smaller firms are lagging, often just using AI features embedded in vendor tools.”

Pleasant’s essential argument is that AI can be great, and many companies are using it, but it needs to go well beyond merely drafting emails or writing blogs. Before it can genuinely fulfil its potential, it needs to be truly integrated into business workflows, changing how companies operate at a fundamental level.

Meanwhile, the hype is driving a sense of urgency that isn’t always matched with readiness. Companies are deploying without strategy, security, or guardrails.”

“Just last week, my local government launched a legal chatbot,” Pleasant said. “The terms of service were laughable. It basically said, ‘We make no guarantees, and we’re not responsible for any bad info this chatbot gives you.’ That’s the reality. Everyone’s covering themselves legally because mistakes are inevitable.”

“Don’t rush in just because your CEO says, ‘We need AI.’ Make sure you have a thoughtful strategy, solid security, and proper governance in place.”

AI Washing, Governance Gaps and the Risk of Rushing In

The rest of the panel is similarly mixed on AI’s current position on the hype cycle.

“I think it really depends on the type and size of the organisation,” suggested Derrick Kelly, VP of Solutions Enablement at AVI-SPL. “We’ve had successful global deployments with large enterprises, well-planned, phased, and properly supported. On the other hand, we also see organisations just starting to ask basic questions.”

“There’s a lot of opportunity, but that includes both legitimate implementations and people trying to sell AI before companies are ready. When that happens, and deployments go wrong, it can set the industry back. Those stories travel fast.”

Kelly outlined that the successful rollouts AVI-SPL has observed have involved careful planning, stakeholder consultation, and a phased approach, rather than as a knee-jerk reaction to the hype: “But that’s true of any technology; you have to do it right, or the results will suffer.”

Melody Brue, Vice President and Principal Analyst at Moor Insights and Strategy, agrees that organisations currently feel like they are wielding a hammer and looking for a nail, using AI for the sake of it instead of solving real problems.

“When surveys say ’90 per cent of companies use AI,’ that might just mean someone used ChatGPT once to ask a question, not that AI is driving business transformation,” Brue said.

We’re in a typical hype cycle: overpromise, underdeliver, followed by a massive investment wave. CEOs are saying this must be a multi-trillion-dollar market; otherwise, the hundreds of billions already invested have gone to waste.”

However, Brue asserted that while the expectations might be out of control, that doesn’t mean there isn’t still incredible substance underpinning it.

“There’s hype, but it’s also pushing investment, innovation, and momentum,” Brue highlighted. “Microsoft’s Work Trend Index talks about ‘frontier firms’ leading the way. Those efforts need support from the top and champions at every level.”

“I just came back from ServiceNow’s Knowledge conference, and I got the same message there. Yes, jobs will be replaced, but others will be created. People saying “AI will take all our jobs” are underestimating human adaptability. AI only knows what humans have taught it. If people stop feeding it ideas, it runs out of steam.”

“Still, there’s a lot of ‘AI-washing’ going on. I saw a post recently where someone shared an iPhone screen protector marketed as ‘AI-optimised.’ That’s when you know the hype is out of control.”

What IT Leaders Could Really Gain From AI Next: Value Over Vanity

For Craig Durr, Chief Analyst and Founder of The Collab Collective, this stage of the hype cycle marks the “trough of despair,” as in the gulf between expectations and reality.

However, he also offered a positive angle around AI’s progress: “The most successful AI I see is what I call ‘boring AI’, used in places like contact centres. You can measure things like time-to-resolution or outcomes that impact real business metrics. Zendesk, for instance, is starting to base licensing on outcome-based performance.”

Kelly agreed that the long-term value lies in total customer experience.

“AI can help unify everything from logistics and reporting to real-time translation in UC&C,” he expanded. “As tools mature and investment continues, we’ll see deeper integration. Agentic AI is the next buzzword, sure, but pairing that with ‘boring’ automation, like moving a customer interaction through the full process without manual intervention, can be powerful.”

AI should assist humans, not replace them. Companies that plan with that in mind will see the biggest gains.”

Durr stressed that these types of deployments, whether in customer experience and the contact centre or in the business world more broadly, often come from pressure at the top: “CEO says, ‘What’s our AI plan?’ CTO says, ‘We’ve got it,’ and CFO says, ‘What’s the ROI?’ The teams that find success are solving boring, measurable problems.”

Brue agreed, drawing parallels with what Bill McDermott calls “soul-crushing work”. “That’s where we’ll first see impact, eliminating repetitive tasks so humans can focus on higher-value work,” she elaborated.

“Eventually, AI will be embedded in everything, like online access or SaaS models,” Durr continued. “The pressure to justify it will fade. But for now, yes, we’re still in the hype phase, like that absurd ‘AI-optimised’ iPhone case.”

“To sum up: yes, use AI, but do it properly,” Pleasant bookended the conversation. “Make sure you have strategy, security, and guardrails. Don’t just jump in because you’re afraid of being left behind or just because your neighbour’s doing it.”

“And don’t just slap an ‘AI’ label on everything!” Brue added.

Agentic AIArtificial IntelligenceCorporate FinanceDigital TransformationProductivityUCaaSUser Experience

Brands mentioned in this article.

Featured

Share This Post