AI customer trust: Why consumers embrace AI but don’t fully trust brands to use it

A shopper asks an AI assistant which of four different dishwashers will fit her kitchen. She reads the answer and buys the one it recommends. 

Two weeks later, she encounters AI again but this time while dealing with an insurance claim. When she asks to speak to a person, she’s redirected to another chatbot. What felt helpful while shopping now feels like a barrier.

Same customer, same technology, opposite experiences. In one moment, AI helps her make a decision. In the other, it stands between her and the help she wants.

Consumers are increasingly comfortable using AI on their own terms. But that comfort doesn’t automatically extend to brands using AI on their behalf. AI customer trust depends on whether customers understand how AI is being used, see value from it, and retain control over the experience.

Our research makes that tension clear. Forsta’s 2026 Retail consumer study found that among shoppers who used AI over the holiday season, 58% bought something because of its recommendation. Asked how far they trust retailers to use AI responsibly, only 22% said a lot or quite a bit.

Consumers are embracing AI faster than they’re granting brands permission to use it on their behalf.  The aforementioned customer’s disappointment has identifiable causes. Our research validates that earning AI customer trust requires giving people enough transparency, value, and control to feel confident about what your brand does with AI.

Consumers have embraced AI but on their own terms

AI has become part of everyday life, and Forsta’s internal research shows how routine: 43% of consumers use it for research and finding information, 34% for personal tasks, 25% for work, and 20% while shopping.

The behavior underneath those figures is consistent. One in three U.S. shoppers used AI over the recent holiday season, and among those who did, 60% came back to it repeatedly rather than trying it once, per Forsta’s Retail Consumer Study. Returning to a tool signals it earned a place in how someone shops.

What people use it for stays narrow. 87% turned to AI primarily for gift ideas rather than price or product comparisons, and fewer than one in ten completed a purchase through an AI platform in any category.

That openness shouldn’t be mistaken for unconditional enthusiasm. Younger consumers may be among AI’s most active adopters, but their skepticism is rising just as quickly. Bentley University-Gallup research found that 47% of adults aged 18 to 29 now believe AI does more harm than good, up 11 percentage points from 2025 and from just 30% in 2023.

That was the largest year-over-year increase of any age group. For brands, the takeaway is that familiarity with AI doesn’t automatically translate into trust. Younger consumers may be comfortable using the technology, but that experience can also make them more discerning about where, when, and how brands deploy it.

Consumer AI adoption is outpacing trust

Forsta’s 2025 State of CX report, drawn from 4,000 consumers across the U.S. and UK, found 48% of U.S. consumers and 45% of UK consumers open to AI-led customer experiences when it means faster service. Only one in five feel very comfortable dealing with AI on its own. 

Consumers don’t automatically trust brands to use AI responsibly, and the disparity between open to and comfortable with is where AI customer service earns a relationship or quietly costs one.

On personalization specifically, Forsta’s internal research finds consumers divided almost evenly in three ways: comfortable, uncomfortable, and undecided. The undecided third is the group worth designing for, because their position is still moving.

Comfort drops further as consequences rise. Close to half remain uncomfortable with AI-generated financial or insurance recommendations, where being wrong costs real money.

The published retail data shows the same caution among people already using the technology. Asked how far they trust retailers to use AI responsibly, 22% of AI users said a lot or quite a bit, and 39% said a little or not at all. The largest group, at 40%, landed on somewhat, which is a hedge rather than a verdict.

Underneath sits the data question. 69% of U.S. consumers will share personal information in exchange for a better experience, while 19% trust brands to handle that information responsibly. The UK figures track closely, at 64% and 17%.

Four questions shaping AI customer trust

Consumers are asking questions brands haven’t fully answered. Four come up repeatedly in the research, and each maps to a decision someone inside the business has already made without telling the customer.

How is my data being used?

Privacy ranked as the top concern among shoppers who used AI, though most of that concern sat in the moderate range rather than the extreme, which makes it addressable. What people want is visibility into how their information feeds the system. Forsta’s State of CX research found 43% of U.S. consumers trust a brand more when it discloses AI use openly, which is a low-cost benefit to give them.

How are decisions being made?

Accuracy and transparency each drew 18% of AI users describing themselves as very or extremely concerned. The issue underneath both is explainability. A recommendation that arrives without reasoning asks the customer to take it on faith, and faith is exactly what’s in short supply. Showing why a product surfaced often does more for confidence than improving the recommendation itself.

Who is accountable?

Customers hold the brand responsible, not the model. When an AI customer support interaction gives someone the wrong answer, nobody files it under vendor error. That exposure grows as AI speaks for the brand more often: nearly eight in ten AI users say they value AI recommendations as much as or more than advice from a retailer. Governance and oversight are what keep that influence from becoming liability.

Can I trust the outcome?

Bias drew the lowest concern scores of the four, which probably understates it, since people can only report bias they noticed. Some did. One respondent described AI that seemed “biased toward certain vendors and products.” Consumers need to believe a recommendation is accurate and serves them rather than the brand paying for placement.

Personalization requires a stronger value exchange

Consumers expect clear benefits before sharing their data, and the State of CX research shows they will trade when the return is legible. 69% of consumers say they will share personal information for a better experience. That willingness is conditional on the exchange being visible, which is where most programs fall down. The data goes in, the benefit stays vague, and the customer concludes the transaction ran one way.

What counts as a meaningful return is unglamorous but ultra-meaningful:

• Support that resolves faster because the agent already has the context
• Recommendations that reflect what someone actually bought instead of what the segment bought
• Loyalty rewards that arrive without being chased
• An experience that picks up where the last interaction left off
• Offers and communications that reflect what someone actually needs right now 

The research shows what happens when AI in customer experience misses that mark. Shoppers praised the technology for speed and convenience, then described missed context, irrelevant results, and generic suggestions as the main friction. Forsta’s own summary of the finding is that AI is winning on utility and lagging on relevance. Generic personalization is worse than none, because it proves the data was collected and reveals it wasn’t used.

Around 30% of consumers in both markets say they would consider switching brands for more personalized experiences, and nearly one in five already have. The value exchange isn’t a philosophical position. It’s a retention number.

AI customer trust and expectations vary by generation

One AI strategy won’t resonate with every customer, and the research splits cleanly enough to plan around. AI use in shopping skews younger, though the retail study notes adoption expanding across age groups rather than concentrating at one end.

The sharper divide is over data. The State of CX research found 49% of Gen Z willing to share personal information against 18% of Boomers, with the Gen Z willingness explicitly conditional on seeing value returned. The same research describes Boomers disengaging from digital-first retail altogether.

The obvious read is that younger customers are the easy audience. The data says otherwise. 71% of Gen Z and 68% of Millennials have walked away from a retail purchase over a poor experience. They adopt faster, share more, and leave sooner. Their openness is a rolling assessment rather than a settled preference, which means a brand earning it has to keep earning it.

Older customers hold the opposite position and are more consistent about it. Emphasis on security, preference for a person on complex or high-stakes issues, and more scepticism toward AI-driven recommendations. That’s a higher bar to clear once and a more durable relationship afterward.

Neither group wants the same AI, and neither is served well by the average of the two.

Human interaction remains essential

AI should enhance human experiences rather than replace them, and consumers are consistent about where the line falls. The State of CX research found that even in highly digitized sectors, many people actively seek human contact when resolving complex or high-stakes issues. Only one in five feel very comfortable dealing with AI on its own.

The pattern in the retail data points the same way. AI use concentrates in high-consideration categories where decisions are complex and confidence matters, and it stops short of the transaction. Among AI users, 5-8% completed an instant purchase through an AI platform in any category. People use AI to think through a decision and then take the decision themselves.

That instinct intensifies as stakes rise. A gift recommendation carries a low cost of being wrong. A mortgage application, a claims decision, a diagnosis, or a complaint that has already failed once carries a high one. Financial services, insurance, healthcare, and customer support are where the demand for a person is loudest, and where routing someone into an AI-only path does the most damage.

The practical version is unremarkable. Automate the volume. Make it easy to reach a human when the stakes are high. Make the handover to a person easy to find rather than buried behind three menu levels.

How brands can build trust in AI customer experiences

Building AI customer trustcomes down to three things: transparency, value, and control. Tell customers when AI is involved and how their data is being used. Make the benefit of sharing that data visible in the experience they receive. And give them meaningful control, including an easy route to a person when they want one.

Sixty-three percent of U.S. consumers will leave after one or two bad experiences, so responsible AI design is a standard applied to every interaction rather than a policy published once.

Consumers have already made AI part of how they make decisions. They use it, they trust its recommendations as much as a brand’s, and will keep doing so on their own terms. What they haven’t decided is whether to extend that confidence to the companies deploying it, and that question is still open in every sector represented in this research.

The organizations that earn customer trust in AI will be the ones that treat transparency, control, and access to a person as design requirements rather than as concessions. Start with disclosure, since it costs the least and moves trust the most.

Speak to an expert about building trust in your customer experience program today.