Artificial intelligence is becoming part of everyday digital commerce, but consumers are drawing a clear line between low-risk convenience and high-stakes decisions. Experian's 2026 Identity and Fraud Report finds that people are increasingly willing to use AI to shop and transact online, while trust falls sharply when the technology is asked to handle travel purchases, financial services or other consequential decisions. The divide is creating a new challenge for businesses: proving that both humans and AI agents can be trusted.

Experian's 11th annual Identity and Fraud Report highlights a tension at the center of the AI economy. Consumers are experimenting with AI for digital transactions, but adoption remains closely tied to perceived risk, while businesses are simultaneously dealing with faster and more sophisticated forms of fraud.

According to Experian's research, 31% of consumers have already used AI tools to shop or transact online, while another 23% say they would consider doing so. Yet only 21% are comfortable relying on AI to complete travel-related purchases, and just 17% are comfortable using AI for financial services decisions.

The disparity suggests that consumers may accept AI as a shopping assistant before accepting it as an autonomous decision-maker.

That distinction is likely to become increasingly important as agentic AI moves beyond generating recommendations and begins completing tasks. AI agents can search for products, compare options, initiate transactions and, in some cases, execute purchases on a user's behalf.

Experian has been building toward that model with Experian Agent Trust, a framework designed around a Know Your Agent, or KYA, approach. The company says the technology can connect a verified consumer with the AI agent acting on that person's behalf while evaluating identity, consent, intent and risk.

The trust problem is not theoretical. Fraud is already increasing as criminals gain access to tools that can automate attacks, generate convincing content and impersonate legitimate users.

The Federal Trade Commission said consumers reported $15.9 billion in fraud losses in 2025, up from more than $12 billion in 2024. The agency received roughly 3 million fraud reports in 2025, although reported losses represent only a portion of total consumer harm because many incidents go unreported.

Experian's report points to several AI-related threats businesses are watching closely. AI-generated phishing attacks ranked as the leading concern, cited by 53% of organizations, followed by AI-assisted first-party fraud at 51%, document forgery at 45%, automated bot attacks at 40% and deepfake voice scams at 37%.

The response is increasingly AI-driven as well. Experian found that 80% of businesses are already using machine learning or generative AI in fraud-management environments. The company's 2026 report also says 93% of organizations express confidence in adaptive risk-based authentication, while 91% have confidence in emerging Know Your Agent capabilities as agentic commerce develops.

That creates an arms race of sorts. AI can help attackers produce fraud at greater scale, but the same technology can help financial institutions, retailers and digital platforms identify suspicious behavior faster.

The more difficult problem may be preserving customer experience while adding those controls.

Consumers appear willing to cooperate when security measures are clearly connected to preventing fraud. Experian found that 84% are willing to complete additional security steps when necessary to prevent fraud, while 71% say it is important for businesses to recognize them accurately online. Nearly half say they have greater trust in organizations that can identify them without repeatedly asking for authentication.

That finding has implications for identity technology and customer experience design. Security does not necessarily have to mean more friction. In a mature identity architecture, businesses can use device, behavioral and contextual signals to make risk decisions without constantly interrupting legitimate customers.

This is where the market is moving beyond conventional authentication toward continuous identity intelligence.

Traditional approaches often focus on verifying a person at a specific moment, such as login or payment. Agentic commerce introduces a different problem: a business may interact with an AI system rather than directly with the human who authorized the transaction.

Experian's Agent Trust framework addresses that distinction through human-to-agent binding, linking a verified individual, device and AI agent. Its platform also uses an Agent Trust Token intended to provide real-time signals around identity, consent and fraud risk, while an Agent Registry evaluates agent behavior over time.

The concept reflects a broader industry shift. As AI agents become participants in digital commerce, businesses will need to establish not only who a customer is, but also whether the software acting on that customer's behalf is authorized to perform a specific action.

For financial services, this is especially consequential. Only 17% of consumers in Experian's survey were comfortable allowing AI to make financial-services decisions, suggesting that the industry faces a much higher trust threshold than sectors involving lower-value purchases.

At the same time, businesses cannot simply avoid AI. Fraud prevention teams are already using machine learning and generative AI, while customer-facing organizations are experimenting with AI-assisted service, personalization and autonomous commerce.

The result is a new digital trust equation in which identity, authorization, fraud detection and AI governance increasingly overlap.

For enterprise technology leaders, that means identity infrastructure may become an essential part of the AI stack rather than a separate security function. Platforms that can verify users, bind agents to authorized principals and evaluate risk continuously could help organizations adopt agentic workflows without treating every autonomous action as an unknown transaction.

Experian's research ultimately points to a market where convenience alone will not determine AI adoption. Trust will.

Market Landscape

The digital identity market is expanding from conventional authentication toward continuous identity, behavioral risk assessment and agent-aware trust infrastructure.

Large technology ecosystems including Microsoft, Google, Salesforce and Amazon are investing heavily in AI agents and enterprise automation. As those systems gain the ability to execute tasks rather than merely provide recommendations, identity and authorization become increasingly important.

At the same time, fraud platforms are adding AI-based detection, while identity vendors are moving toward adaptive authentication and continuous risk evaluation. Experian is positioning Agent Trust in the emerging gap between consumer identity and autonomous software activity.

The competitive landscape will likely include identity providers, fraud-prevention platforms, payment networks and enterprise AI companies. Differentiation may come down to how effectively each can connect human identity, agent authorization, transaction intent and real-time risk without adding unnecessary friction.

Strategic Outlook

The most important change may be that businesses will increasingly need to treat an AI agent as a distinct participant in a digital transaction.

That does not mean trusting every agent equally. It means establishing a verifiable relationship between the human or organization that authorized the agent, the permissions granted to it and the behavior it exhibits during a transaction.

For marketers and commerce teams, this evolution could also reshape personalization and customer engagement. An AI agent acting for a consumer may eventually negotiate offers, compare products and complete purchases, changing how brands think about identity, attribution and customer relationships.

The technology challenge is substantial, but the market signal is clear: agentic commerce will require a trust layer that is designed for machine-mediated transactions.

Top Insights

• Experian's research shows consumers are embracing AI for online activity but remain cautious about high-stakes decisions, creating a significant trust barrier for agentic commerce.
• Reported U.S. fraud losses reached $15.9 billion in 2025, increasing the urgency for identity and risk technologies capable of addressing AI-enhanced attacks.
• Businesses are using AI defensively as well as offensively, with 80% reporting machine learning or generative AI use in fraud-management environments.
• Experian Agent Trust introduces human-to-agent binding and Know Your Agent concepts designed to connect identity, authorization, intent and risk in autonomous transactions.
• For enterprise technology teams, identity is becoming part of the AI infrastructure stack as software agents increasingly act on behalf of customers and employees.

Get in touch with our MarTech Experts