How Do Customers Really Feel About Using AI?

Customer sentiment around AI is shifting, but not always in the way organizations expect. While capabilities continue to improve, customer confidence is moving more slowly; creating a gap between technological progress and lived experience.

This tension is reflected in the latest findings of The State of Customer Service and CX (2026), so we spoke to Shep Hyken to find out more.

Customers Don’t Trust AI to Work Consistently Across Different Brands

On paper, AI in customer service is advancing rapidly. Resolution rates are improving, more customers are successfully using tools like ChatGPT, and a majority are noticing better performance year-on-year. Yet resistance remains stubbornly persistent.

The core issue is not simply whether AI works, but whether customers trust it to work consistently across different brands and scenarios. This inconsistency creates hesitation, particularly when customers are unsure what experience they will get next time around.

Shep Hyken, CAO (Chief Amazement Officer) at Shepard Presentations LLC — Shep Hyken

“Not every company is evolving at the same speed as others, so customers are still left thinking, “I wonder what the experience is going to be like next time?”, so until they actually have numerous interactions with a brand over a prolonged period, they won’t feel confident of the service they’ll receive.

And this is likely going to be interwoven with negative experiences with other brands, who may even still be trying to recoup their investment in legacy systems.” – Shep Hyken

Despite the research showing that 53% of customers have noticed an improvement in AI customer service technologies over the past year, and 50% of customers now saying they have successfully resolved a customer service issue using AI or ChatGPT, customers are still trying to get answers in a very fragmented CX landscape.

This ultimately means that perception is shaped less by what AI can do and more by how well various organizations are delivering AI-enhanced service.

Customers Need to Know a Human Will Be Available When They Need Them

What’s also increasingly clear is that AI alone should not be the end goal. Instead, CX leaders should be striving for a blended approach, where AI handles initial interactions and humans remain available when needed.

Why? This hybrid model is essential to maintaining trust and ensuring customer confidence does not deteriorate when automated systems reach their limits.

“When we’re talking about AI strategy, this should never mean AI only! A human should always be ready to take over when needed.

When a company provides that, so it’s easy and intuitive to transition to a human when needed, it’s a winning combination.

I often tell my clients, let’s go AI first but then have the human backup ready to go. Never make it hard to get to a human – otherwise you lose all trust with the customer!” – Shep Hyken

After all, customers are not necessarily resistant to AI, but they are highly sensitive to friction and lack of control.

Simply Making AI Available Is Not Enough

And it’s not just about designing the right experience; it’s about actively encouraging customers to engage with it too!

Quite frankly, those who want to see real shifts in AI adoption amongst their customers will need to rethink their approach to encouraging behavioural change. Simply making AI available is not enough!

Customers need to be guided, incentivized, and reassured that these service options are effective. This is particularly relevant in contexts where customers still default to phone-based support even when digital options are available.

To understand the misplaced confidence that technology alone will fix deep-rooted operational challenges, read our article: Is “AI Optimism” Compromising Your CX Strategy?

Adoption Rates Will Continue to Climb So Long as Brands Focus on the Right Things

Looking ahead, expectations for AI in customer service are high, but current progress remains incremental.

While customer preference for phone-based support is still dominant, long-term sentiment suggests a gradual shift toward digital-first engagement. However, this shift is not simply about technology adoption, it is about usability, and perceived reliability!

“It’s important to remember that whilst a year is a long time in how technology evolves so quickly, it’s a short time for how humans typically take to something new.

We see it happening when we ask the question, “Would you rather call for customer support or go online?”

Currently 7 out of 10 customers prefer to make a phone call when there’s a problem, and that number has been pretty consistent over the last few years, but I think we’re going to start to see a big leap soon.

Maybe not so much in the next year, but I believe in the next few years we’re going to see that number increase. You’ll see more people adopting it because it’s easier, it’s better, and it’s more intuitive. Also, because more smart companies will be pushing it WITH the human backup firmly in place.

Unfortunately, there will still be those that get it wrong. As those that decide to push this to a point where they make it hard to get to a human will lose the relationship battle with the customer.

Whilst the winners will be those that create a great AI-first experience with backup! So, when the customer needs additional help, relationship-building skills kick in, and the query is managed human-to-human once more.” – Shep Hyken

Customer Sentiment Towards AI Is Best Understood as Cautious Optimism

Overall, customer sentiment towards AI in CX is best understood as cautious optimism. Whilst capability and successful resolution rates are improving, trust is still being shaped by inconsistency across brands and the quality of hybrid experiences.

But as the latest research suggests, customers are not rejecting AI! Rather, they are evaluating it through the lens of reliability, ease of use, and the ability to escalate to human support when needed.

Therefore, until those expectations are consistently met, perception and adoption rates will likely continue to lag behind technological progress.