Why 2026 Is the Year Customer Expectations Outran Customer Experience
Part 1 of "The Experience Gap" series
I’ve sat through a lot of quarterly business reviews where someone puts up a slide showing the CX budget going up and to the right, nods around the room, meeting adjourned. Nobody asks the harder question: is any of this money actually reaching the customer? This year, the data finally forces the question.
Verint just published its State of Customer Experience 2026 report, 5,000 U.S. consumers surveyed, and buried in it is a number that should ruin a few executives’ week. For the first time in five years of running this research, a majority of consumers, 51%, say businesses fall short when they actually need help. Not in some edge case. Not on a bad Tuesday. As the baseline.
Let that sit for a second. The default experience is now “you let me down.”
It’s Not That Service Got Worse. It’s That We Moved the Goalposts on Ourselves.
Here’s the part that’s easy to miss if you only skim the headline stat: this isn’t really a story about companies getting worse at service. Most of them are trying harder than ever. It’s that customers stopped grading on a curve.
Forty-two percent say their expectations have gone up this year, versus just 19% back in 2024. More than double, in two years. And why wouldn’t they? Nobody’s comparing your support chat to the support chat they had last time. They’re comparing it to the Instacart delivery that texted them before they even wondered where their groceries were. To the assistant that flagged a problem before they noticed it themselves. To the one company, somewhere, that actually fixed their issue on the first try.
That’s the bar now. Most organizations built their service model for the old one.
The Investment Is Real. I Just Don’t Think It’s Going Where People Assume.
To be fair to the leadership teams catching flak here, the money is showing up. Generative AI spend in customer service is growing at something close to a 28% compound annual rate this year, and analysts expect the market to nearly triple by decade’s end. This isn’t a story of companies burying their heads in the sand.
But look at what that spend is actually producing. Gartner doesn’t expect agentic AI to autonomously resolve most common service issues until 2029, three years out. Today, self-service resolution sits around 14%. And here’s the stat that stopped me: nearly two in three customers say they wish companies would use less AI in support, not more.
So we’re spending like it’s working and hearing from customers that, mostly, it isn’t yet. Both things are true at once, and I think that’s the real headline of 2026.
The Uncomfortable Bit: Nobody Told the Agents
This is the stat I keep coming back to. Roughly one in five frontline agents, 21%, actually has generative AI tools in hand while they’re on a call or a chat with a customer. One in five.
Which means for every press release about an AI rollout, there’s a much larger, much quieter reality: the people doing the actual work of closing the gap in real time mostly haven’t been given the tools yet. The budget got approved. The pilot got announced. The rep on line three is still working off the old playbook.
I don’t think that’s a technology failure, honestly. It’s a sequencing failure, buying the capability before doing the unglamorous work of putting it in front of the people who need it, inside the workflow they already live in.
Not everyone’s stuck there, though. Bank of America just expanded EricaAssist, its AI agent that now sits alongside more than 18,000 customer service reps during live calls, surfacing account context in under three seconds without breaking the flow of conversation, and shaving nearly a minute off average call time in the process. Small detail, but it’s the right one: the AI is helping the human, not replacing the interaction. We’ll come back to that model in a few weeks, because I think it’s closer to what “done right” actually looks like than most of what’s been announced so far.
Why I Think This Is Really a Trust Story
It would be tidy to file all this under “technology adoption problem” and move on. I don’t think that’s it, or at least not the whole of it.
When expectations climb this fast while actual service stalls, what you’re watching is trust erode in slow motion. Every letdown lowers the starting point for the next interaction. And here’s the part that should actually worry you: customers increasingly don’t complain when this happens. They just quietly do less business with you, or leave without a word. That silent-churn problem is coming later in this series, because I’d argue it’s the most underpriced risk sitting inside most CX dashboards right now.
For today, sit with the headline number. Most of your customers think you’re falling short when it counts, not for lack of effort, but because the gap between what’s been promised at the strategy level and what actually lands with them has never been this wide.
What’s Coming Next
Week 2: Where the AI investment dollars are actually going — and why spend and value are decoupling
Week 3: The frontline enablement problem, and what Bank of America and others are doing differently
Week 4: Why CX maturity, not technology adoption, is the real dividing line between leaders and laggards
Week 5: Silent churn — why fewer customers complain, and why that should worry you more, not less
Week 6: A practical playbook for closing the gap
My bet: the organizations that win the next phase of CX won’t be the ones with the most AI. They’ll be the ones who actually close the distance between what they’ve bought and what the customer on the other end of the line feels.
Next week: following the money behind the CX AI boom, and why bigger budgets haven’t translated into better outcomes.