← All posts

November 4, 2026 · 2 min read

43% aren't sure an AI agent's request is legitimate. Can your customers trust yours?

New SEON research found that 43% of U.S. adults are not confident they can tell whether an AI agent's request is legitimate, and the top trust builders were human support and the ability to reverse mistakes. Here is how a small business can make its own AI agent easy to trust.

What happens when your AI agent sends a customer a message, and they are not sure it really came from you?

That question is closer than it sounds. A new survey from SEON, a fraud prevention company, asked 2,026 U.S. adults about AI agents in early October. 68% already use an AI agent at least once a week for a personal task. Yet 43% are not confident they could tell whether an agent's request for information or account access is legitimate. Among Boomers, that number rises to 61%.

The survey also asked what would build more trust. The top answers were human support if something goes wrong (50%), the ability to reverse mistakes (42%), and strong identity verification (39%).

A separate American Bankers Association survey, conducted by Morning Consult, points the same way. 80% of consumers say a bank should be able to tell whether it is dealing with a customer or the customer's AI agent, and 19% say an AI agent has already asked them for login credentials or card numbers.

Why this matters for a small business

If you use an AI agent to send reminders, collect documents, or follow up on invoices, your customers are getting messages from software. Many of them cannot easily tell a real request from a scam. The trust you have built over years lives in that moment. People are not asking for perfect AI. They are asking for a person they can reach, and a way to fix a mistake.

That is why I keep coming back to one simple idea: automate the process, not the relationship.

What I would do this week

  • Read every message your AI agent sends and make sure none of them ask for passwords, card numbers, or bank logins. Then tell your customers you will never ask for those by message.
  • Add one clear line to every agent message with a way to reach a real person on your team.
  • Write down how you would undo an agent's mistake, like a wrong reminder or a wrong invoice, and name the person on your team who owns that fix.

My takeaway: Your customers will trust your AI agent as much as they trust the human standing behind it. Make that human easy to find.

Sources

  1. New SEON Survey Finds U.S. Consumers Are Embracing AI Agents, but 65% Would Blame the AI Company if One Commits Fraud · SEON (vendor research, November 2, 2026)
  2. ABA survey: U.S. consumers wary of AI agents handling their money, bank account information · ABA Banking Journal, American Bankers Association (October 27, 2026)