From AI adoption to AI trust: What CX leaders must get right now

Executive summary

Artificial intelligence (AI) is rapidly becoming embedded across customer experience (CX)—from service and personalization to marketing, analytics and creative. While AI adoption is accelerating, industry data shows that customer trust has not kept pace. The 2025 Edelman Trust Barometer (The AI Imperative) highlights a critical reality for CX leaders: trust is now the primary determinant of whether AI-driven experiences are embraced or resisted.

This article draws on insights from the Edelman research and practitioner discussions with the CMA CX Council and members of the CMA’s AI Committee to outline what it takes to build trust in AI-powered customer experiences. It argues that AI success in CX depends less on the sophistication of the technology and more on how intentionally it is designed into the customer journey. Clear customer benefit, transparency, human choice, and peer influence emerge as the strongest trust drivers supported by responsible and ethical AI practices, including appropriate governance, fairness, accountability and human oversight —while poorly governed automation, lack of authenticity, and invisible efficiency plays actively undermine confidence.

For CX leaders, the implication is clear: AI can no longer be treated as a behind-the-scenes capability. It is a frontline brand experience that demands deliberate design, stronger governance, ethical and responsible AI practices, and renewed focus on human judgment. Organizations that embed AI thoughtfully—earning trust one interaction at a time—will be better positioned to scale adoption, deepen loyalty, and sustain credibility in an AI-powered future.

Artificial intelligence is no longer edging its way into customer experience—it is already embedded. Across industries, AI is being introduced into service models, marketing, personalization, analytics and creative workflows. In many organizations, its use is not just encouraged but required, and employees are increasingly expected to demonstrate AI proficiency as part of their roles. Yet as adoption accelerates, one thing has not kept pace: trust.

When we shared findings from the 2025 Edelman Trust Barometer: Trust and Artificial Intelligence at a Crossroads with the CMA CX Council, the tension was unmistakable. AI is becoming unavoidable, but customer comfort remains uneven—and in some markets, actively resistant. That gap between deployment and trust is where CX leaders now sit, accountable not only for whether AI works, but for how it feels. This is no longer a theoretical concern. Customers experience AI not as a strategy or a roadmap, but as a series of moments: a chatbot interaction, a translated message, an ad, a recommendation, or a response generated in real time. Each interaction quietly answers a question that customers are increasingly asking themselves—is this helping me, or helping the company? They know the difference.

AI is now a frontline CX experience

The Edelman data makes one thing abundantly clear: trust is the conduit to AI adoption. People who trust AI are dramatically more likely to engage with advanced use cases, from financial decision-making to healthcare management and major purchases. Conversely, lack of trust doesn’t just slow adoption—it actively shapes skepticism toward brands that feel opaque, self-serving, or dismissive of human judgment. For CX leaders, this shifts AI from a back-office efficiency lever to a brand-defining capability. AI now sits squarely within the experience layer, influencing loyalty, credibility and long-term relationships. What emerged from the Council discussion was not a debate about whether AI belongs in CX—it does—but a shared recognition that AI must be designed to earn trust, not assume it. That means embedding responsible AI principles - including transparency, accountability, fairness and human oversight into every customer facing experience, not treating them as compliance exercises after deployment. 

Designing for trust requires more than good intentions

Across the research and practitioner discussion, four elements consistently surfaced as foundational to trustworthy AI-driven customer experiences. Importantly, they do not work in isolation. They must be designed together, deliberately and visibly.

First, trust is built when customer benefit is immediate and obvious. Customers are far more open to AI when they can clearly see how it reduces effort, increases clarity, or helps them move forward faster. When AI is introduced primarily to deflect service, cut costs, or optimize internal throughput—without a perceivable improvement to the customer—confidence erodes quickly. Consumers sense when efficiency is prioritized over experience, even if they don’t articulate it explicitly.

Second, transparency has become non-negotiable. Customers do not object to AI nearly as much as they object to feeling misled. Whether through AI-generated chat responses, advertising creative, or automated recommendations, opacity breeds suspicion. Transparency doesn’t require deep technical explanations; it requires acknowledgement. Customers increasingly expect organizations to be open about how AI is used, how decisions are made, and how personal information is handled responsibly. Simply signaling when AI is involved, what role it plays, and where its limitations are goes a long way toward reducing uncertainty and restoring credibility.

Third, human choice remains essential. Chatbots, in particular, surfaced as a recurring stress point. While customers appreciate AI for simple, transactional needs, tolerance drops sharply when nuance, emotion or judgment enters the interaction. People want the option to escalate to a human—not as a last resort, but as an embedded choice. When AI feels forced, trust collapses. When it feels assistive, trust grows.

Together, these practices reinforce the principles of ethical and responsible AI by ensuring technology augments human decision-making rather than replacing it where empathy, judgement or accountability matter most.

Finally, peer influence acts as a powerful multiplier. The Edelman findings reinforced what many CX leaders see anecdotally: people trust “someone like me” far more than CEOs, institutions, or experts when it comes to AI. Peer stories, testimonials and shared experiences accelerate confidence not only in AI itself but in the surrounding experience design. This applies equally to employees and customers. Trust spreads horizontally, not top-down.

Where CX leaders are losing trust—often unintentionally

The Council conversation surfaced several practical examples of how trust breaks down when AI is poorly implemented. Translation and localization emerged as a particularly vivid example. AI has made it easier than ever to scale content across languages, but scale without representation is quickly recognized. In markets like Quebec, where linguistic nuance and cultural specificity carry deep importance, AI-generated translations that miss the mark signal exclusion rather than efficiency. The cost savings may be real, but the experience communicates that the customer was not fully considered.

Another tension appears in the growing reliance on AI-generated insights. While AI is powerful at synthesizing large volumes of data, it can confidently surface outdated or context-poor conclusions if not governed carefully. Without human experience applied as a filter, teams risk making CX decisions based on insight that looks authoritative but isn’t relevant. Human oversight remains essential to identify bias, validate recommendations, and ensure AI outputs are fair, accurate and appropriate for customer context. Authenticity and content provenance also surfaced as critical trust concerns. As generative AI makes deepfakes and synthetic media more accessible, the bar for quality assurance has risen dramatically. Review layers including licensing checks, IP clarity, scene-by-scene scrutiny are essential safeguards of credibility. AI does not reduce the need for rigor; it amplifies it.

Trust requires new operational muscle

Perhaps the most important takeaway for CX leaders is that trustworthy AI experiences cannot be bolted onto existing operating models. They require new muscle. Governance must be proactive rather than reactive, built in partnership with legal, privacy and compliance teams from the outset. Responsible AI governance should also establish clear accountability for AI enabled decisions, monitor model performance over time, and ensure customer privacy and regulatory requirements are consistently met.

Education must be tailored, recognizing that leaders, managers and frontline employees experience AI differently—and have different anxieties around its impact. Change management can no longer be an afterthought, particularly as AI use becomes mandated rather than optional. From a strategic standpoint, over-reliance on a single AI tool or partner introduces new forms of risk that CX leaders must account for. In short, AI introduces speed. CX leaders now have to introduce intention.

The CX leader’s moment

AI adoption will continue to increase. Market pressure, executive mandates and competitive dynamics ensure it. But trust will not follow automatically. For CX leaders, this moment represents a shift from experimenting with AI to designing it responsibly and ethically—with customers, not just systems, at the centre. The most advanced AI-driven experience is not the one with the most automation or intelligence layered in. It is the one that customers feel comfortable returning to, confident that their interests have not been abstracted away. Organizations that consistently demonstrate responsible and ethical AI through the principles in the Canadian Marketing Code of Ethics and Standards including transparency, fairness, accountability, data privacy and security, and meaningful human oversight will be best positioned to earn lasting customer confidence. 

AI may be unavoidable—but trust still must be earned, one experience at a time.

Authors:
Jeanette Kennedy, Global Marketing Lead, Microsoft Canada
Ankit Wadhawan, Director, Next Best Action, RBC