In a recent LinkedIn post, product strategist Melissa Perri discusses the critical distinction between User Interface (UI) and User Experience (UX) as it pertains to generative AI tools, cautioning business leaders against over-reliance on these technologies without human oversight.
Perri highlights that while AI tools excel at creating visually appealing interfaces, they fundamentally lack the understanding of user needs that underpins true UX. She states:
“These models are trained to produce what looks right. They have never met your customer, sat in a usability test, or watched someone get stuck. So they optimize for what is plausible and polished, which is the surface. That is UI. Whether a real person can actually use the thing is UX, and that is exactly what they cannot see.”
As Melissa Perri elaborates, this blind spot in AI models, which are trained on vast datasets to generate plausible outputs, leads to several predictable issues when used for product development or website creation.
The Homogenization and Usability Gaps of AI-Generated Design
One significant concern raised by Perri is the tendency for AI to produce designs that lack originality. According to Perri, when left to their own devices, AI models often default to the same standardized patterns, resulting in websites and applications that look virtually identical.
Furthermore, Melissa Perri points out that AI can construct user flows that, while appearing polished, are not practical for real users. She notes that these flows might lack essential elements like error states or present information in an illogical sequence, all while maintaining a superficially professional appearance.
“It builds flows nobody can actually use. You get a screen that looks clean and professional, but the model never stops to ask whether a person could move through it. It will hand you a form with no error states, or ask for information in an order no human would expect, and it looks great the whole time.”
Perri also emphasizes the potential for AI to fabricate information or features. She recounts instances where AI generated policies the company did not offer or proposed services outside their scope, underscoring the need for vigilant human review.
The Critical Role of Human Expertise and Skepticism
Despite these challenges, Melissa Perri is clear that these AI tools remain valuable. However, she strongly advises users to approach them with caution and maintain a critical perspective.
“None of this means the tools are not useful. We shipped a whole site with one. It means you cannot hand your judgment to a model that has never met your user.”
In Perri’s view, the most crucial element in leveraging AI effectively is the presence of individuals with deep UX expertise. These professionals are essential for guiding AI tools and identifying their errors. She expresses concern over the trend of companies prioritizing junior staff to accelerate AI adoption while potentially diminishing the role of experienced professionals.
“This is exactly why having people with real expertise matters so much. They are the ones who can steer these tools and catch what they get wrong,” Perri writes. “So it worries me to see companies laying off their most skilled and senior people while keeping juniors to move faster with AI.”
Perri concludes by reiterating that while AI can generate the visual elements of a product, it cannot ascertain their effectiveness for end-users. That responsibility, she asserts, remains firmly with human product teams.
📝 About This Content
This article is based on insights shared by Melissa Perri on LinkedIn.
📅 Originally posted on July 30, 2026 | View original post on LinkedIn →