For success in AI, avoid the ‘efficiency trap’— and focus on trust instead

If customers don’t trust how companies deploy AI, they’ll walk away. If employees don’t trust it, they’ll disengage. If enterprises don’t trust their AI providers, they won’t adopt. A recent global KPMG study found that while two-thirds of people now use AI regularly, fewer than half say they’re willing to trust it.

I see this playing out every day in customer experience. Deployed thoughtfully, AI can supercharge a company’s ability to understand customers and build connections that drive loyalty. But when the primary objective is to cut costs and reduce headcount, AI can just as easily alienate customers and destroy the trust that took years to build.

Trust and understanding build on each other. When customers trust you, they share feedback, engage more deeply, and give you the signals that reveal what they actually need. That understanding helps you build better products and experiences, earn more trust, and generate more insights.

Trust is high stakes

Think about the last time you tried to get help from a company using AI. Maybe you were bounced between chatbots that couldn’t solve your problem, or you got stuck in a rigid workflow with no way out, desperately hunting for an empathetic human who could actually help.

You’re not alone. Our research shows nearly one in five consumers who have used AI for customer service saw zero benefit. That’s a failure rate almost four times higher than the AI failure rate in general. Half of consumers worry AI will prevent them from ever reaching a human. Concern about how companies use personal data for automation is up 10% from last year.

Every frustrating interaction customers and employees have with businesses hurts trust.

The stakes are enormous. It’s estimated that bad customer experience places $3 trillion in global revenue at risk every year. Roughly half of consumers cut spending with a brand after a single bad interaction.

The efficiency trap

When AI conversations in boardrooms start and end with cost reduction, failure is likely.

I see two kinds of AI innovation in the market: science projects and pragmatic AI. Pragmatic AI solves real problems, makes people’s lives easier, and builds connections. Science projects impress in demos but frustrate in deployment.

The cost-cutting obsession is a trap, and the savings may prove illusory. Gartner predicts that by 2030, the cost per resolution for generative AI in customer service will exceed that of offshore human agents.

Every competitor has access to the same models, cloud infrastructure, and automation playbooks. When efficiency is the only goal, companies build a commodity.

Worse, an AI strategy obsessed with cutting costs often produces impersonal, frustrating experiences that drive customers away. You save money on the front end and lose it, along with the customer relationship, on the back end.

What AI winners look like

Companies that get this right stop thinking of AI as only a cost play and start seeing it as a way to better understand their customers, build deeper connections, and deliver business outcomes.

Take customer feedback programs, which are the bedrock of customer experience. When AI is used to make surveys smarter, more conversational, and respond to vague and unactionable feedback, something remarkable happens. When customers share more feedback, quality improves, and the business can take more meaningful action. Organizations are seeing more than 80% of customers engage, and the feedback coming through is twice as good as before.

The real prize for companies is when they act on this feedback. We’ve integrated agentic AI into surveys powered by an organization’s policies and runbooks. When a customer provides feedback, the agent can take action to close the loop immediately. For companies, this is translating into huge jumps in customer satisfaction and tens of millions in returns.

In healthcare, we have seen systems deploy AI programming that does an excellent job of remembering who patients are – recording their history and special concerns. This doesn’t replace doctors and nurses. It informs them and frees them from administrative busywork so they can focus on the human relationship that patients value.

In market research, companies are using AI, including synthetic market research, to test ideas, understand demographics in hours instead of months, and consolidate platforms. The increased speed is great, but what really matters is what’s being done with the insights: launching products that respond to what customers want while competitors are still working to understand their customers’ needs.

In each case, the technology is almost beside the point. What matters is that these companies understand their customers better. And that understanding compounds into trust and outcomes.

Trust is the great differentiator

Trust is one of the greatest and enduring differentiators in business.

Companies that earn trust get better data. Better data means better AI interactions. Better AI means better experiences. Better experiences mean more trust. It’s a flywheel, and with every cycle, the leaders pull further ahead.

Meanwhile, companies that deploy AI without adequately considering its impact on people find themselves stuck. Their AI doesn’t improve because customers won’t engage with it. Their customer relationships erode because every interaction feels like talking to a machine that neither knows nor cares about them.

Think about where we’re heading. Five years from now, there’s a legitimate question about how often we’ll interact with websites through browsers versus having conversations with agents.

In that agentic world, the experience delivered to the user will differentiate every provider.

The organizations that understand human connection – personalization and the ability to make customers feel heard and seen – will win. Because they are the companies that figured out that AI works best when it helps you understand, respond, and build connections with people, not just process them.