Companies are racing to adopt AI, but its success depends on confident adoption, and adoption depends on trust. Building that trust starts with recognizing the sentiment toward AI among customers and employees, and knowing what to do about it.
To help, my colleague Dr. Ben Granger, Chief Workplace Psychologist at Qualtrics, and I combined insights from four independent studies spanning over 100,000 consumers and employees to understand their interactions with and perceptions of AI. Here's what we found.
Customers need to trust AI is there to help, not to stand in the way
Consumer comfort with AI has fallen from 54% in 2024 to 48% today, and even reached a low of 43% in 2025. Only 29% of customers say they trust organizations to use AI responsibly, while misuse of personal data is now consumers' top concern when companies use AI to automate interactions.
These interactions fail when customers don’t trust an AI agent to act in their best interests, and when it stands between them and the help they actually need. Speed without usefulness isn't a benefit. It tells the customer the organization chose cost over their experience.
I experienced this firsthand when refinancing my house. Your typical, standard process, including the 149 lender spam calls I got on the first day. Then my credit report came back with no mention of my mortgage. I went to the agency’s site to sort it out, but was relegated to a chatbot: “File a dispute online! It’s free,” it assured me, “the agency will get back to you in about a month.”
Mortgage rates move fast; I didn’t want to wait that long. But all the chatbot could do was assure me it “understood my frustration.” I tapped out.
Employee adoption of AI depends on earning trust too
The employee picture is more nuanced, but the underlying cause is similar: trust determines whether adoption grows. And right now, most organizations need to do more to earn it.
While daily organizational usage is up, optimism has plateaued. Our research shows only 57% of employees are involved in decisions about how AI will affect their work, 54% say their organization has clear usage guidelines, and 50% have received formal training. All three metrics have moved fewer than two points year-over-year.
When people don't know why AI is being deployed, what the rules are, or whether they can trust the output, uncertainty drives avoidance. Sanctioned tools go unused, and employees route around them with personal accounts and ungoverned applications they’re more familiar with. This introduces data risks, inaccurate claims, and compliance exposure the organization was trying to prevent.
The data shows that exposure helps. Employees who use AI tools daily or weekly are over 40 percentage points more optimistic than those who use them monthly or less. The more people actually use these tools in ways that work, the better they feel about them. But that flywheel only spins when people are equipped and invited in, not when they feel forced.
Trust is the key to AI adoption
We interpret these findings through a model of organizational trust built on three recurring dimensions:
- Competence (can the organization do the job?)
- Integrity (does it act consistently in alignment with its stated values?)
- Benevolence (is the organization’s implied motive genuinely in my interest?)
Most organizations clear the first two bars. 70% of global employees rate their senior leaders as competent and 67% rate their leaders as having integrity. But ratings of benevolence—whether leaders are acting beyond the short-term profit motive—fall to 55%. This lack of benevolence is where AI deployments fail, because employees aren't convinced the organization deployed them for their best interests.
For organizations that can address this gap, the value compounds. In our benchmark of high-performing organizations, an elite group of companies that outperform on both employee experience and financial outcomes, the single largest differentiator between the top 5% and the middling majority is a 19-point difference in employee trust in senior leadership. As we know from extensive client work connecting employee and consumer insights, the trust among employees inside the organization strongly affects consumer trust in the organization.
What it looks like when organizations get trust right
The most successful AI rollouts we’ve studied—those built on trust that drive organic adoption—share several principles:
Show the benefit, specifically
The clearer the answer to "what's in it for me," for employees and customers alike, the higher the adoption and comfort. Organization-specific training and visible use cases drive usage better than broad mandates.
Frequency drives comfort (and vice versa)
Although it’s helpful to continually educate employees on the value of AI, it’s also helpful to nudge people to use the tools. When you do this, mindset often follows, and a virtuous cycle is created.
Give people meaningful control
People who are given objectives and leeway adopt faster and more durably than those given rigid instructions. Pair that autonomy with enablement—prompt libraries, success stories, and guidelines instill confidence rather than constrain it.
Protect human-centered moments
Not every interaction is an efficiency problem. High-emotion, high-stakes moments are where human connection is key. Consider starting with back-of-house operations to enhance person-to-person interactions and expand once trust is established.
Design with feedback in mind
Measure the AI experiences you introduce, and continually use customer and employee feedback to inform AI rollouts and change communications. We saw a healthcare organization in our research that acted on feedback to cut down a claims process from two hours to two minutes with AI tools to bolster the patient experience.
Build governance and guardrails
Autonomy works best when paired with training and enablement, prompt libraries, and a cross-organizational group of AI stewards that provide the guidance and guardrails that instill confidence.
Purpose-built AI earns trust
Recall the earlier chatbot incident. While in the moment, I was frustrated with that particular AI tool, yet the very next morning, I used an internal AI tool to gut check whether this story made a useful point. It gave me direct, specific feedback and let me decide what to do with it. Same person, 24 hours apart—but a completely different experience, because the second tool had a clear, bounded job, and I was in control.
Purpose-built AI that’s designed with your organization’s context, guidelines, and operational knowledge performs more reliably, is more trustworthy, and is easier to adopt. Employees understand what it's for. Customers interact with something that demonstrably understands their situation and can do something about it.
That's the path from only 29% of consumers trusting an organization’s use of AI to something meaningfully higher. Not faster automation, but AI that improves the overall experience for the people it's meant to serve.
Matt Evans leads the employee experience product science team at Qualtrics. Dr. Ben Granger is Chief Workplace Psychologist at Qualtrics XM Institute.
