From Hype to Habit: Experience-Led Ways to Design AI Products That Actually Get Used
- Nichola Wolfe
- Nov 21
- 7 min read
Most of what we read about AI sits at two extremes.
On one side, there is the technical world: model tuning, data pipelines, inference speeds and engineering decisions that only a small percentage of people truly understand.
On the other side, there are big-picture narratives. AI will revolutionise everything. AI will transform the workplace. AI will change how we live, learn and work. Interesting, but often too broad to help anyone build something practical.
The space in the middle is where most AI products succeed or fail. It's the part where a team moves from an idea to something people trust, understand and choose to use. That middle space is shaped by experience, based on how real people behave, how they make sense of something new and how the product fits into the wider context of their day.
Across my work in AI product experience and adoption, I’ve seen the same pattern repeat; The model can be strong. The business case can be clear. The ambition can be exciting. But unless you design for the lived experience of the user, the product rarely takes hold.
Below are practical, experience-led considerations that help bridge the gap between technical build and real uptake.
1. Start with research and keep it close throughout the work
Good AI products begin with a clear problem, and user research is the only reliable way to find it. Spend time with the people who will eventually use the product. Listen to how they describe their work. Observe where the friction sits. These early conversations protect the team from building something that feels impressive but solves little.
Research shouldn’t end once the problem is defined. It should continue through workflow mapping, early concept testing and post-launch refinement. Each stage offers different insight: what people do today, where the product should sit, how early designs land and how behaviour shifts once the AI is introduced.
When a team keeps research close, decisions stay grounded. When they don’t, assumptions quietly move in and shape the direction.
2. Design the product in context, not in isolation
Every user already has an ecosystem of tools, habits and workarounds before your AI product arrives. Understanding this ecosystem is essential. It prevents duplication, shows you what the product needs to connect with and highlights any friction that might get in the way.
This is not only good for experience — it’s good for the business. It protects teams from investing in something that already exists elsewhere, or from creating a feature that competes with another internal tool.
A clear view of the ecosystem helps the team design a product that complements what people already do instead of interrupting it.
3. Place the product where it will make a meaningful difference
Workflow design isn’t about finding the exact pixel where a feature should sit. It’s about understanding where decisions are made, where people hesitate and where support genuinely helps. AI products often work best when they reduce effort or uncertainty at a specific point in the day.
If a product forces people into a separate space or asks them to remember another step, it becomes one more thing to think about. If it supports them in a moment they already recognise as challenging or time-consuming, the value is felt immediately.
Clear workflow insight helps the team identify these moments. Without it, the product risks becoming something people only use when they remember, rather than something that naturally supports their work.
4. Treat trust as a design requirement
Trust is not something users “eventually develop.” It’s something you design for from the start. People want to know what the AI can and cannot do. They want clarity around the data behind it. They want to understand why a recommendation appears, even at a basic level. They want reassurance that mistakes are expected and not career-limiting.
But trust goes further than the product itself. When it comes to AI products aimed at employees, people also want to understand the purpose behind it. If the only message they hear is “this will make you more productive,” it’s natural for them to wonder what that productivity is for and whether that shift reduces their value in the long run. This is where leadership, strong change management and communication matter. A product team cannot build trust alone. Users need a clear sense of how the AI fits into the future of their role, what opportunities it opens and how it strengthens the organisation rather than replacing people within it.
Without that foundation, even well-designed products struggle. The team ends up pushing uphill because the broader narrative is missing. With it, people approach the product with more confidence and curiosity, and the experience work has room to take effect.
5. Give Design a strategic seat at the table
From my experience over the last couple of years, I’ve noticed a pattern in how AI product teams are formed. The technical capability gets built out quickly, and then one designer is added with a blended “UX/UI” remit, often well into when the project starts and sometimes with only partial capacity. When this happens, design becomes stretched, and the product naturally leans towards what the technology can do rather than how people will actually interact with it.
It’s important to create balance on the team. Designers bring far more than visual polish. They shape the interaction model, help teams understand how decisions are made and imagine how the AI will be experienced long before anything is built. When they are involved early, they help connect the technical possibilities to the realities of human behaviour.
Prototypes become much more valuable in this environment. They allow teams to test ideas quickly, observe where people hesitate and identify what needs to change. When designers and engineers work together, the product evolves through a cycle of evidence, observation and refinement, rather than assumption or technical momentum.
Design is a strategic capability in AI. When it sits alongside engineering and product, the work becomes more balanced, more grounded and far easier for users to adopt.
6. Define meaningful measures that move beyond speed and productivity
AI products are often positioned around doing things faster. Shorter cycles. Fewer manual steps. While useful, these signals don’t explain why the product matters or what changes because of that added speed.
Meaningful measurement sits a few layers above productivity. If a task becomes quicker, what does that free someone up to do? If a workflow becomes lighter, how does that improve the quality of decisions or the pace of progress? If a team saves time, where does that time go and what does it enable?
When metrics connect to real outcomes, they help everyone understand the value the product is meant to deliver. They show how the work feeds into something larger and how that moves the organisation forward.
AI products feel more purposeful when measurement reflects progress, not just pace.
7. Design the adoption journey with the same care as the product
AI products require people to change how they work. This deserves deliberate attention. Provide space for people to explore the product without pressure. Give them opportunities to ask questions and to practise using the AI before it becomes part of their everyday tasks. Support managers so they understand how the product fits into expectations and workload. Above all else, don't let a comms campaign do the heavy lifting. Consider the different type of adopters in your customer or employee ecosystem and create adoption initiatives that are relevant to them and reflect their level of capability.
When teams treat adoption as a separate, thoughtful piece of work, the transition feels lighter. Users arrive at launch with familiarity instead of uncertainty.
8. Shape the roadmap using what you learn after launch
AI products evolve based on how people use them. The first release is rarely the version that delivers the most value. Pay attention to how people interact with the product. Look for the moments they ignore suggestions, where they override recommendations or where they spend more time than expected.
These early signals help shape the next iteration. They reveal missing information, unclear steps or insights that need more context. They help the team refine the product so it becomes genuinely useful.
A roadmap built on real usage is more reliable than one created months in advance.
What leaders can do to support good AI product work
Leaders play a significant role in helping Product Owners build AI products that land. The pace of AI development is high, and it’s natural to feel pressure to deliver quickly. But speed without purpose can create its own challenges. Teams rush into solving the wrong problem. Users experience yet another tool without understanding how it fits into their day. Change fatigue builds, and the product struggles to gain traction.
Support from leadership doesn’t mean slowing down. It means creating the right conditions for clarity. Giving teams the time and space to understand why the product matters. Helping them connect the work to a meaningful organisational outcome. Encouraging them to involve users early rather than pushing straight into delivery. Removing barriers that make iteration difficult. Providing access to the skills and data needed to build something well.
When leaders set this tone, teams move with more confidence and less rework, and users feel part of a thoughtful shift rather than another wave of change. It creates momentum that is grounded rather than forced.
Closing thoughts
AI continues to move quickly, but the fundamentals of good product design haven’t changed. Real understanding of the user. A clear view of the workflow. A thoughtful approach to trust. Metrics that reflect real outcomes rather than surface-level speed. A product shaped by ongoing learning rather than static assumptions.
When these elements come together, AI becomes something people choose to use, not something they feel obliged to try.
That’s the shift from hype to habit, and it’s where long-term value begins. At its core, it reflects the classic experience principle: make products people want, instead of trying to make people want a product.