Trust is the currency of conversational AI in higher education

“Trust is the currency” has become a mantra for us at the Conversation Design Institute. And the more we work with universities deploying conversational AI, the more convinced I am that it captures the single most important truth about building AI that actually works for people. Not automation rates. Not cost savings. Not containment metrics. Trust.

Education changes lives. It empowers individuals to break cycles of poverty, builds tolerance between cultures, and provides the foundation for economic development. The around 250 million students enrolled at 21,000 universities worldwide, along with governments, families and individuals, spend more than $5.6 trillion a year on education and training. The stakes could hardly be higher.

Just 36 per cent of Americans now express significant confidence in higher education, down from 57 per cent less than a decade ago. In the UK, more than 40 per cent of universities will have budget deficits, with some institutions threatened with collapse. The situation is described in Shitij Kapur’s Triangle of Sadness, in which aspiring but anxious students burdened with debt combine with stretched governments ambivalent about universities’ public good with beleaguered staff caught in the middle. Universities are now faced with competition to provide more services while cutting costs.

In this context, many institutions have turned to AI as a solution. But here’s what they’re discovering: without trust, AI makes things worse, not better.

Every customer interaction is an exercise in trust. When someone reaches out for help, they’re placing their confidence in you—trusting that you’ll understand them, respect their time, and genuinely try to solve their problem. It’s always been true in human customer service. What’s changed is that we’re now asking Conversational AI to carry that responsibility.

The uncomfortable truth is most AI assistants are failing spectacularly at earning trust. They’re technically functional but emotionally vacant. They process queries without understanding context. They provide answers without anticipating needs.

The result? Customers don’t trust them. They learn to bypass them, looking for the “talk to a human” button before they’ve even tried the chatbot. Every abandoned conversation represents a small erosion of the relationship with the people you serve.

For universities already facing a crisis of confidence, this erosion is something they simply can’t afford.

Jerry Michalski has spent decades studying why institutions struggle with trust. His insight is that most of our systems—from education to governance to customer service—have been designed from an assumption that people can’t be trusted.

This explains why so many university chatbots fail the way they do. They’re designed from mistrust: assuming students will try to game the system, that queries need to be verified before being helped, that the default response should be caution rather than care. The result is AI that treats every student like a potential problem to be managed rather than a person to be supported. When we design systems from trust the outcomes are radically different.

If trust matters, it matters exponentially more for students. The relationship between a university and its students is fundamentally different from a typical customer relationship. Students aren’t just buying a product—they’re entrusting their futures, their wellbeing, and often their mental health to an institution.

Today’s students expect fast, personalised, 24/7 support—not just in learning but across their entire educational experience. When a university’s chatbot asks them to repeat information they’ve already provided, or offers generic responses to deeply personal questions, or fails to recognise when a simple query is actually a cry for help—that student learns something important. They learn that the institution doesn’t really know them. That perhaps it doesn’t really care.

This is the mistrust signal. And students receive it loud and clear.

It means designing assistants that assume good intent—that treat a student’s question as legitimate before requiring them to prove it. It means building systems that share information generously rather than gatekeeping it. It means creating AI that makes the institution vulnerable in small ways—acknowledging uncertainty, admitting limitations, offering to connect students with humans when the stakes are high—because that vulnerability signals that the institution trusts students enough to be honest with them.

So how do we build AI assistants that earn trust rather than erode it? The answer lies not in better technology but in better design—specifically, human-centred conversation design that prioritises relationships over transactions.

At the Conversation Design Institute, we believe great AI assistants aren’t just built—they’re cultivated. They emerge as the natural outcome of aligned people, shared purpose, responsible design, and strong organisational systems—always in service of the student.

Our Standards Framework evaluates AI capability through four lenses: Mindset (the beliefs that shape how teams treat people), Skillset (the competencies to design and deliver safe, human-centred AI), Culture (the norms and values that sustain empathy at scale), and Systems (the processes and tools that make trust repeatable). We’re not just designing for trust. We’re helping universities design from trust.

Universities are, at their best, are places where trust is extended to young people, where they are given the opportunity to grow, to fail, to discover who they are and who they might become. That trust has always been the foundation of education’s transformative power.

The universities that will thrive in the Conversational AI era won’t be those with the most sophisticated technology. They’ll be those that use AI to deepen their relationships with students—to make every student feel known, valued, and supported at scale.

Because in the end, trust really is the currency of Conversational AI. And like any currency, you can either invest in building it—or watch it slowly drain away.

For more on our approach to Conversational AI in higher education, visit

https://conversationdesigninstitute.com/conversational-ai-for-higher-education

You can take our free Conversational AI Maturity Assessment here

https://scorecard.conversationdesigninstitute.com/education

Or read our insights into Conversational AI in Higher education here

https://www.conversationdesigninstitute.com/conversational-ai-for-higher-education/insights