
Nearly half of consumers have already walked away from a brand, or spent less with it, because of how it handles their data in AI systems. In our State of Digital Trust research, a survey of 11,000 consumers across seven markets, 47% globally said they had taken an action with a direct revenue consequence for exactly that reason. That is not a future risk to manage. It is money that has already moved.Â
Most companies still file this under privacy and compliance, where the job is to meet a standard and avoid a fine. That framing made sense when the main risk was regulatory. It makes much less sense now that customers are the ones enforcing the rules, with their attention and their spend. Trust has become something the business feels in retention and revenue, and it needs to be read there too.Â
AI changed who enforces the rulesÂ
AI has made customer data more useful and more visible at the same time. Businesses can now personalise and act on that data in ways that weren’t possible a few years ago. Customers have noticed, and they have grown quicker to react when a company’s use of their data crosses a line they didn’t agree to. 74% say they would stop using a service after their data is misused.Â
Regulators still set obligations, and those obligations still matter. But meeting them only tells a company whether it cleared the legal bar. It says nothing about whether customers are comfortable with what happens after they click “accept.” Those are different questions, and only one of them shows up in the numbers leadership watches.Â
Transparency has a price, and customers are naming itÂ
The clearest sign that trust carries commercial value is that customers will pay for it. In our research, 52% said they would pay more for brands that are transparent about how AI uses their data, at an average premium of 7%. People are putting a real number against data practices they can understand and feel in control of. When a behaviour shows up as a price, it belongs in commercial planning, not just the compliance report.Â
The counterintuitive part: informed customers are the comfortable onesÂ
The instinct is that the more people learn about how their data is used, the warier they become. Our data points the other way. Among consumers who understand their privacy rights, 53% are comfortable with AI-driven personalisation. Among those who don’t understand them, that figure drops to 19%.Â
That gap matters, because 71% of consumers globally find AI-driven personalisation intrusive. The discomfort isn’t really about personalisation itself. It tracks with how well people grasp what is happening to their data. A retailer that explains, in plain language, what its recommendation engine does with a shopping history can get a warmer reception for the same feature than a competitor that stays quiet.Â
What measuring trust looks like in practiceÂ
The data shows trust now moves revenue. The harder question is what a business does about it.Â
It’s tempting to treat trust as a feeling nobody can measure, so nobody tries. Given what it now moves, that’s a hard position to defend. Trust itself may be difficult to score, but its effects on behaviour are already sitting in systems companies run every day.Â
This doesn’t mean inventing a “trust score.” The more useful signals already sit in the business. After a company makes its data practices clearer, the things to watch are the ones it already tracks: retention, churn, opt-in rates by segment, and the volume of data-access requests and complaints. For teams rolling out AI features, adoption is another read: are customers switching the features on, or leaving them alone?Â
These signals also expose a gap that companies rarely see in themselves. A business can believe it is being transparent because it publishes a data notice. If customers still hesitate to share data or turn on AI features, they have read the situation differently. The disclosure existed; the understanding didn’t.Â
None of this belongs to the privacy team alone. The people who own retention and revenue need to see it too.Â
The stakes rise when AI starts asking for moreÂ
This gets sharper as AI moves from analysing data to acting on it. AI agents that do things on someone’s behalf need deeper access to personal data and to the accounts people use. Many want to grant that access one request at a time: 23% globally will allow it only if they can approve each individual request, and just 8% are fully comfortable letting an AI assistant reach their data with no conditions at all.Â
For any business betting on these tools, that number sets a ceiling. A more capable system is worth only as much as the access customers are willing to give it. If they hold that access back, the capability sits idle. Earning permission has to happen before the feature ships, not after.Â
Put trust where the business can see itÂ
AI has raised both the value of customer data and the cost of losing people’s confidence in how it’s used. That confidence already shows up in the numbers: whether customers stay, and whether they use what a company builds next. The businesses that handle the next phase of AI well will be the ones that know how much trust they have banked before they ask customers for more. The ones that keep trust filed under compliance will be reading the wrong number when it starts to move.Â



