Europe's AI ambitions outpace consumer trust, report says

Okta has published its Customer Identity Trends 2026 report, which points to a widening gap in Europe between corporate AI ambitions and consumers' willingness to let AI act on their behalf.

Fewer than one in five European organisations are on schedule with AI transformation work, even though three-quarters of European developers and senior leaders say their organisation has introduced or plans to introduce AI into customer journeys.

Compliance and governance were cited as the main reasons for delays. Developers also pointed to identity systems that cannot support the access patterns used by AI agents.

The research suggests businesses face a separate challenge as they introduce more autonomous services. While 40% of European consumers said they were willing to delegate some decisions to AI agents, only 7% were ready for AI to take actions on their behalf.

The findings come as European businesses face pressure to adopt AI while meeting new regulatory demands. The EU AI Act has sharpened the focus on how companies govern AI systems and manage user trust.

Trust gap

Across the wider global sample, the report found broad acceptance of AI in digital life, but less support for unsupervised action. It found that 75% of consumers welcome AI in digital life, while 71% would walk away if an AI took unauthorised actions on their behalf.

The research also points to the commercial cost of poor customer experiences. Globally, 74% of consumers would abandon a site or app if they were forced to create an account before seeing key information, and 72% would leave if login or verification took too long.

Security expectations also remain high. Some 89% said they value clear control over their data and how it is used, while 93% said trust depends on strong security without difficult access.

Those figures suggest companies are being pushed in two directions at once: to make digital journeys simpler while also proving that customer data and permissions are being handled properly.

Gareth Davies, Chief Product Officer at Auth0, said the low level of consumer readiness for autonomous AI should not be surprising.

"It is perfectly reasonable that only seven percent of European consumers are ready to let an AI act on their behalf. After all, handing an autonomous agent your credit card without any guardrails is a rather bold financial move. But the promise of frictionless agentic commerce cannot be ignored either. Just as we built consumer trust in the early days of online shopping through robust fraud protection, we must now build it for the agentic era by establishing verifiable human intent. To close this confidence gap and safely capture the high-converting revenue these agents bring, businesses need modern identity infrastructure that dictates exactly who an agent represents and what it is authorised to do before a transaction ever occurs," Davies said.

Project delays

Material accompanying the report says almost 60% of organisations are moving to introduce action-taking agents, yet 80% of those projects are running behind schedule. It also suggests senior leaders and developers see different causes for those delays.

According to the figures, 59% of business leaders blame hold-ups on compliance, security review or governance issues. Developers were twice as likely as senior leaders to say their current identity infrastructure cannot support agent access patterns, with the research putting that figure at 17.8% versus 9.0%.

That difference points to a divide between executive ambition and technical readiness. It also highlights identity management as a practical barrier to deployment rather than a background IT issue.

The report argues that businesses using an external customer identity platform are more likely to keep AI rollouts on schedule than those using legacy or internally built systems. It says 28.3% of organisations using an external platform report staying on track, compared with 18.1% using other approaches.

It also says developers using external identity platforms are more likely to have AI features embedded in production, at 34.4% compared with 22.3% for others. The same group was less likely to report losing customers because of security friction.

Jas Sagoo, Head of International Auth0 Strategy, said identity controls sit at the centre of the issue.

"Our latest CIAM research makes one thing clear: consumers aren't handing AI a blank cheque. Businesses in EMEA need to earn people's trust if they want AI agents to play a bigger role in their lives, and there's a clear gap between what businesses want AI to do and how much decision-making consumers are willing to hand over. Identity has a central role here: recognising the customer and respecting the permissions they have given. The EU AI Act is setting the guardrails in this process, but real trust is won in the user experience. The businesses that lead will use AI to eliminate friction while keeping people firmly in control of their identities and decisions," Sagoo said.