The Dirty Little Secret About AI in Contact Centers

As an industry, we've never stopped chasing the elixir of faster response times, lower costs and round-the-clock service. Wouldn't you be chasing it if your CXO kept calling you the "cost center?"

And now another shiny object emerges: AI everything. But for all its dazzling potential, AI is eroding the trust in the very experiences it purports to improve.

The Trust Disconnect

It's not entirely our fault. We attend trade shows, sit through keynotes, and walk away ready to automate every interaction in the name of speed and scale. However, automation without credibility creates a type of friction that damages trust.

When customers don't trust the automation, whether it's a chatbot or the AI agent pretending to be a person, they revolt. They opt out. They escalate. They repeat themselves. They call back. Suddenly, your "efficiency gain" is nothing more than the newest bottleneck.

I've always found it strange how easily we forget that we're customers, too, when designing workflows and solutions. Am I the only one who has angry-typed the word "AGENT" in all caps? Or spoken it out loud in my best all-caps voice, hoping to trigger a live transfer? We don't do that because we don't hate automation. We do it when faced with a system that feels unhelpful or, worse, indifferent. Our customers don't need the automation to be perfect. They need it to inspire trust.

Customers Don't Want Speed. They Want Support.

According to the Financial Times, companies are eager to embed generative AI across the customer journey, but many are struggling to convert data into meaningful, human-centered value. Customers are looking for the true promise of AI through experiences that feel intelligent, conversational, and context aware.  

Not All Friction Is Bad

Don't get me wrong, there's room for a little friction in CX. In fact, Harvard Business Review points out that the rush toward frictionless AI experiences can backfire. In emotionally sensitive environments, small, intentional moments of friction can enhance trust, support better decision-making, and preserve customer agency.

"Good friction," as HBR calls it, gives people time to reflect, understand, and stay in control. It allows them to ask, "Do I want this?" before a system says, "Here it is."

For example, if a customer is trying to cancel a service, you don't have to lose the rebuttals. Gently remind them of what they will lose if they cancel (i.e., playlists, downloaded content, watch history) and offer them the option to pause the subscription or to set a reminder to reconsider canceling next month. But also, a clear path to cancel.

The Path Forward

At the heart of every automated exchange is a human expectation that isn't tied to response time or technical precision. Trust is built on clarity, consistency and care. So, if you're deploying or planning AI in your contact center, here's a simple framework to help build trust in the system before your customers head to their nearest exit:

• Label bots clearly. It is okay to lead with "This is an automated assistant." Your customers are no dummies, and it serves no one to pretend there's a person on the other side.
• State the task for the end user. "I can help schedule, provide account updates…" or whatever the AI is performing should be clearly stated to set the expectation and remove possible ambiguity.  
• Explain the escalation alternatives to the customer. It lets them know if things get complex, there's a transfer to be had.  
• And speaking of a transfer to be had, seriously, give them the option to get out of there. Some people just are uncomfortable with the tech. Give them the option to type 'agent' and get routed to a person. Agency builds trust.
• Don't just ask for customer feedback and ignore it. Use it for a more intelligent design.

In Summary

Trust is the currency of automation. Without it, even the most advanced systems will fail to connect with the people they're designed to serve.

When customers feel confused, controlled, or dismissed by AI, they disengage, or worse, they escalate or leave. Churn comes at a cost in the form of time, money and brand equity.  

Let's be intentional about how we implement AI rather than chase what is technically possible. Let's prioritize what earns trust by building systems that are believable, respectful and human-centric.