Only 6% of consumers want AI to make final decisions and act entirely on their behalf, according to Okta’s 2026 Customer Identity Trends Report. Okta published the report and its November 4 analysis globally. 

The report examines consumer attitudes toward AI-driven digital experiences alongside businesses’ readiness to deploy agentic systems.

The research covers consumers and technology leaders in the United States, Canada, the United Kingdom, Germany, Australia, New Zealand and Japan.

The report draws on surveys of 5,250 consumers and 910 developers and senior business leaders across seven markets. The consumer research and organizational findings cover different respondent groups.

The findings point to a gap between enterprise adoption plans and consumer willingness to delegate decisions. While 32% of consumers said they would be comfortable allowing AI to handle routine tasks while retaining control over important decisions, another 37% preferred receiving AI-generated suggestions and making decisions themselves. Twenty-five percent wanted little or no AI involvement.

Fifty-two percent of consumers believed AI capabilities in apps and websites were advancing too quickly. Their willingness to delegate varied according to the consequences of an action and whether they could review, approve or reverse what an agent did.

For use cases where the financial or personal consequences of an error are limited, willingness to let AI act on a consumer’s behalf rises above 50% when the person can approve the action first or reverse it afterward.

Enterprise AI Projects Face Deployment and Governance Hurdles

The consumer findings come alongside challenges reported by organizations developing AI-powered customer experiences. In its analysis, Okta said nearly 60% of organizations were introducing action-taking agents, but 80% of these AI initiatives were running behind schedule.

Business leaders most frequently cited compliance, security reviews and AI governance as obstacles. 

Developers were more likely to identify limitations in existing identity infrastructure, which can make it difficult to manage how agents access applications and customer information.

Identity and Permissions Become Part of AI Governance

For chief data officers (CDOs), the findings underline the importance of treating identity and permissions as part of enterprise AI governance. 

Agents that can access customer accounts, retrieve personal information or initiate transactions need clearly defined permissions, traceable activity and controls that allow users to approve or reverse consequential actions.

These requirements become particularly important as organizations connect agents to multiple applications and services. A well-defined AI governance framework should establish the boundaries of an agent’s authority, the data it can access and the circumstances in which human approval is required.