UX design for AI products: Why empathy is the missing ingredient

When a man in Canada reached out to an airline chatbot to ask about bereavement fares after the death of a loved one, the system told him he could book a full-price ticket and then apply for a refund later. That turned out to be false. When the man contacted Air Canada to correct the mistake, the airline initially refused, citing its actual policy. The customer had been misled by its own AI agent.

This isn’t a futuristic dystopia; it happened in 2024.

And it’s not an isolated case. From customer service to healthcare, AI systems are being deployed in more parts of our daily lives, automating decisions that range from helpful to high-stakes. We encounter chatbots, personalized content feeds, voice assistants, predictive UX flows, and AI-powered decision support tools. The promise of AI is convenience, scale, and efficiency. But it often arrives without the nuance that makes UX design for AI products truly work — emotional awareness, ethical reasoning, and contextual understanding..

What’s missing is empathy.

That’s especially true in UX design for AI products, where assumptions can scale into systemic problems.

As the founder of a UX agency that works with health tech and B2B SaaS companies, I’ve seen firsthand how empathy can change the trajectory of a product. In recent years, it’s gone from being a soft skill to a design imperative. The more AI influences our products and experiences, the more essential empathy becomes as a guiding principle.

Why empathy matters more now in UX design for AI products

For years, empathy has been core to UX design. It helps teams step outside their own assumptions and design from a user’s point of view. It’s how we discover not just what users do, but why they do it. At Standard Beagle, we use tools like empathy maps, journey maps, and deep qualitative research to reveal emotional needs that analytics alone can’t uncover.

But the explosion of generative AI has raised the stakes. While AI systems can simulate understanding, they don’t experience emotions. They can recognize sentiment in a message, but they don’t feel concern. They can optimize a recommendation engine, but they don’t know when a user is vulnerable or confused.

As AI scales, these limitations scale with it. A poorly designed human experience might frustrate one person at a time. A poorly designed AI experience can frustrate millions.

According to Forrester’s 2024 U.S. Customer Experience Index, overall customer satisfaction has declined for three years in a row, despite heavy investment in AI automation. A Gartner study found that 64 percent of customers would prefer not to use AI for customer service at all. People don’t trust the bots. And it’s no wonder why.

Why UX design for AI products requires a new mindset

The tools are more powerful, but the risks are bigger too. UX design for AI products forces teams to consider how emotion, trust, and clarity shape outcomes, especially when those outcomes are automated.

What happens when empathy is absent

When AI is designed without empathy, the experience breaks down fast. Many of these failures stem from a lack of strategic focus in UX design for AI products.

• Chatbots become robotic: They fail to recognize distress, get stuck in loops, and offer tone-deaf responses.
• Systems fail to adapt to context: Predictive UX may recommend actions that are irrelevant or even offensive.
• Bias goes unchecked: Algorithms trained on skewed data can reinforce discrimination, especially if marginalized users weren’t part of the research process.
• Trust erodes: People don’t understand how AI makes decisions, and when explanations aren’t provided, confidence collapses.

We’ve seen real examples: Amazon’s AI recruiting tool penalized resumes from women. A healthcare chatbot gave dangerously inappropriate responses. A well-known assistant suggested a child try a viral internet challenge that could have led to injury. These failures weren’t technical glitches. They were failures of design. And they point to a truth UX professionals have always known: Design is never neutral.

Empathy as an engine of trust

Empathy isn’t about being nice. It’s about being deliberate. It’s a method for understanding people’s goals, pain points, and emotional states so we can build experiences that serve them better. In AI systems, that means designing for:

• Transparency: Making sure users understand what the AI is doing and why.
• Fairness: Including diverse user voices in the design and testing process to reduce bias.
• Control: Giving users the ability to override, adjust, or question AI decisions.
• Privacy: Respecting data boundaries and using information ethically.

Empathy is what connects these principles. It puts us in the mindset to ask: How would I feel if this system made a mistake that affected my finances? My health? My child?

Designing with empathy in the AI age

Product teams that want to build responsible, trustworthy AI experiences need to rethink UX design for AI products from the ground up. They need to embed empathy at every stage of development.

• Start with qualitative research: Interviews, contextual inquiry, and diary studies uncover user motivations that aren’t obvious in dashboards.
• Use diverse personas: Represent the full spectrum of your user base, not just the average. Include extreme users. Build in edge cases.
• Co-create with users: Invite users into the design process. Participatory design ensures the solution reflects lived experience.
• Map emotional journeys: Identify where friction, frustration, or anxiety might occur, especially when automation is involved.
• Design for transparency: Show when AI is making decisions, offer clear explanations, and let users opt out.

One example comes from the world of financial services. Chatbots designed to handle customer disputes have been shown to misinterpret claims, offer incorrect guidance, or fail to escalate issues to human agents. These breakdowns weren’t due to technical limitations alone, but to a lack of empathetic design. This kind of oversight is exactly what UX design for AI products must prevent. Had teams conducted deeper contextual research and prototyped with real users, they might have uncovered how emotionally charged and complex these interactions can be and built in clearer feedback loops and escalation paths from the start.

That’s what empathy looks like in practice.

Empathy scales when we practice it, not just preach it

Empathy is often treated as a philosophical idea, something nice to have. Now that AI systems are shaping real outcomes for users, empathy is the guardrail that protects against harm. It’s a way to prevent harm. It’s a way to ensure we’re building tools that people can trust.

And yes, it scales. Not because machines become empathetic, but because we embed empathy into the way we design them.

It starts with our research methods. It shows up in our process documentation. It lives in our stakeholder conversations, our interface decisions, our copywriting. And when teams are diverse, inclusive, and aligned around a user’s lived experience, that empathy becomes systemic.

The tools of AI are powerful. But without empathy, they’re blunt. And nowhere is this more evident than in UX design for AI products, where the interface must translate complex systems into human understanding. As we shape the future of product design, empathy isn’t the opposite of innovation. It’s the path to it.

And it’s the only thing standing between a helpful agent and a harmful one.