Read our ✨ Free UX Assessment Self-Help Guide
Why human expertise still defines great UX
Published on
December 10, 2025
Many organizations continue to expect AI in UX design to be autonomous and generate value without any instructions or oversight. In reality though it’s best to think of AI as a partner capable of adding to human expertise.
Design teams are already embracing this approach. When integrated into a streamlined workflow with clear human oversight, AI strengthens existing standards and reduces friction from concept to release. This makes digital value measurable and consistent. However, the Zeroheight Design Systems Report 2025 shows that more than half of the respondents either avoid AI or have limited plans to adopt it. This skepticism is understandable especially since many of the early tools promised speed but actually introduced more risk and design debt.
But AI is now a part of the design landscape. So, the question is not so much about whether it belongs, but how teams can maintain oversight and accountability.
As technology evolves, so must CMS platforms. Embracing new tools and features will be crucial for staying competitive in the digital landscape. The integration of voice search and chatbots is also expected to become more prevalent.
“Al is not a lead designer. It's a new joiner who learns things fast but still needs plenty of feedback.”
Alexey Novik, UX Department Manager at Coherent Solutions
The human-in-the-loop approach maintains accountability and decision-making firmly with the design team. While automation can efficiently handle repetitive UX tasks such as updating pieces, aligning documentation, or keeping interfaces consistent, human designers maintain ownership over experience quality and user outcomes.
UX design is about interpretation and empathy, areas where human judgment cannot be replaced. While AI streamlines tasks and offers solutions, it lacks the understanding needed to make weigh decisions from the user’s perspective. This is why concerns about AI replacing human designers miss the point. The strength of UX lies in context and empathy, not just efficiency.
Human judgement not only prevents mistakes; it also keeps user-centered thinking throughout all automated interactions. This ensures that AI stays aligned with the strategic intentions behind every design decision.
Clarity and shared context are needed to preserve a productive Human<>AI relationship. This is especially vital in UX design, where intention and nuance guide each interaction.
Effective UX collaboration with AI is dependent upon generating an ongoing, contextual conversation, not isolated commands. If AI can retain context from one interaction to the next, it can generate outputs congruent with design intent. Without context retained, each interaction feels fragmented, increasing the risk of drift or misinterpretation.
This collaboration is a simple loop: the designer describes the purpose, AI proposes, and the team refines the result. These structured feedback loops reduce confusion and make the expectations clear and understood while revealing areas where the AI needs more precise guidance. We've seen first-hand examples of this dynamic in action. Tools promising rapid AI-generated changes can actually slow teams down if the interaction is not clear. For example, small visual adjustments in Figma, with current GenAI capabilities, can take several minutes, while a designer manually adjusting component properties can do the same in seconds.
This is a key UX principle in AI collaboration, assigning the right task to the appropriate partner. Automation excels at highly specific, repetitive work, while designers remain faster and more effective at handling immediate, context-sensitive decisions.
Every teammate needs boundaries, and an AI teammate is no exception.
Clear boundaries determine whether AI can push UX design forward or if it poses a risk. Human<>AI partnerships are successful only if they are based on clear constraints that dictate where automation is allowed to operate, where its data comes from, and what it does in case of uncertainty.
AI-generated UX suggestions need to reference approved, verifiable sources. Having a retrieval mechanism integrated with an internal design knowledge base allows for consistency with pre-defined guidelines and minimizes guesswork while speeding up reviews.
But even strong guardrails can’t make up for poor-quality inputs. Tools trained primarily on aesthetically appealing mockups or untested student projects tend to ignore fundamental UX principles. Attractive interfaces can hide underlying usability gaps. Prescreening training materials against known UX patterns prevents common mistakes from reoccurring.
This mismatch between idealized demos and real-world performance may partly explain declining user interest in some AI-driven design platforms. While marketing demos appear flawless, actual use quickly reveals the lack of meaningful constraints or actionable context.
Embedded constraints give teams confidence that the AI reinforces, rather than replaces, established quality and usability standards.
The strength of human-centered AI is empathy and context, exactly where automation typically fails on its own. While AI can support human insight, it cannot independently replicate user-focused judgment.
For UX teams, using AI for UX design means prioritizing governance and accessibility in design principles. These are then interpreted by human designers and become the benchmarks that determine success.
Organizations that adopt this human-centered AI approach can see measurable benefits. New designers ramp up faster as UX standards are embedded across workflows. Error rates go down, and rework is reduced, allowing teams to focus on actual exploration instead of repetitive fixes.
The payoff shows up in designs that stand up to real-world use.
At The Norm, AI is part of our daily UX practice, but it is never the end goal. It's a tool to help us learn faster and automate routine tasks. But its value depends on people who understand context and intent.
You can see this balance most clearly at the start of a new project, when the team is still building their understanding. Every new or reworked product takes time to build that knowledge, and that learning phase can slow momentum. AI reduces that delay by
managing research and organizing information, so designers can move forward with a clearer foundation.
We’ve seen this firsthand in our internal UX training program, a structured learning initiative designed by our UX practice to onboard newcomers and introduce them to our team workflows, best practices, and design standards using real-world projects. The program is built on the principles of human-centered design (HCD) and prioritizes hands-on experience. Every participant works on these practical projects under the guidance of Senior and Lead designers. The results are then reviewed by the wider team, which turns every exercise into a shared learning experience.
AI is now an important part of that onboarding process. Work that once took days or weeks to complete, such as researching markets, competitors, and user needs, can now be done in just a few hours. AI tools help collect data, analyze customer feedback, review documentation, and suggest discussion points for kickoff meetings. This allows new designers to learn faster and gives us a consistent foundation for evaluating the quality of AI artifacts, tools, and design solutions across our practice.
However, while AI can work around the clock and remove bottlenecks, it still needs supervision. This is primarily because while it can propose directions, it cannot decide which of those directions best supports the user. That responsibility remains with designers.
Human judgement is what keeps designers at the center of every project. There are AI solutions to aggregate and store data, tools for computer vision, and tools that can analyze emotions and feelings. But each of these works in isolation. Humans connect these signals into a coherent understanding. For instance, during research a designer might notice a shift in tone, small contradictions or genuine excitement about a particular idea. These subtle cues can guide the design toward what users actually value.
AI provides the groundwork. Human interpretation determines the outcome. Even the most capable AI design assistant depends on that human interpretation to give its results meaning.
Our UX team recently used AI to extend the Coherent Design System (available in the Figma community) with a custom plugin that maintains brand consistency from concept and mockups all the way through front-end implementation. The result was higher productivity without any loss of quality. Senior designers reviewed every AI-assisted step to ensure alignment with our standards.
The same approach applies to how we work with clients. During live sessions, AI supports collaboration within established design standards and allows us to test and adapt faster while being able to maintain control over quality. This enables our team to brainstorm on the fly and reach outcomes faster, replacing what used to require hours of back-and-forth communication.
Our practical guidelines for incorporating AI into UX are simple:
When AI is guided in this way, it supports human expertise instead of competing or replacing it. Each decision remains intentional. Each result strengthens the user experience.
AI will eventually be integrated into every design process; that goes almost without question. But what will make teams stand out from their competitors is not the sophistication of their automation, but whether the collaboration between designers and AI crafts consistent, human-centered experiences.
Because while AI takes over more technical tasks, from summarizing and analyzing data to generating UIs, compiling layouts, and even scripting interviews, the foundation of great UX will remain built on human psychology and relationships.
Leaving humans entirely out of the equation would lead to AI designing products for AI, not for people. The long-term advantage will be with the teams that maintain empathy and human judgment at the heart of their design.
At The Norm, AI strengthens how we design by supporting human expertise, not replacing it.
We bring that balance to every project we take on.
No. AI can accelerate design tasks, but it lacks the ability to interpret human context or emotion, and it cannot empathize. Human judgement remains essential to interpret users and shape experiences.
AI supports designers with research, documentation, and visual updates. It helps maintain consistency and improves efficiency when guided by human oversight.
AI shortens delivery cycles and lowers error rates. It reinforces standards and lets designers focus on higher-level decisions.
It keeps people responsible for direction and quality. AI assists by extending the designer’s capacity, not by making independent choices.
