JHDD UI UX Report — 2026.09.24

Bilal Skiani’s observation about user trust in AI products highlights an interaction pattern issue deeper than surface-level UI.

The seemingly disparate challenges of hiring new designers, measuring design output, and ensuring user trust in AI products share a common thread: the increasing unreliability of traditional benchmarks. Screen-based metrics no longer capture the full scope of user experience. Resumes focused purely on established expertise may overlook the innovative capacity needed for emergent problems. AI products, while powerful, introduce a critical variance problem that traditional usability models struggle to address. This pattern points to an industry-wide re-evaluation of what constitutes effective design practice and valuable design output.

The variance problem in AI products, as identified by Bilal Skiani, exposes a fundamental shift in user expectations and trust. Users forgive a slow answer if it is predictably slow and explained. What breaks trust is inconsistency. Mainstream opinion often posits that AI will streamline design processes by automating repetitive tasks, allowing designers to focus on higher-level strategic work. This perspective often overlooks the foundational role of predictability in building user trust. An AI tool that generates varied results without clear explanations, regardless of its speed, introduces friction and doubt into the user’s workflow, undermining any efficiency gains. The true design challenge with AI involves designing for transparent and consistent behavior, rather than merely achieving superior outputs. By mid-2028, design systems will explicitly incorporate variance modeling and communication protocols for AI-driven components, shifting the definition of “deterministic” beyond traditional UI elements.

This redefinition of reliability extends to how design talent is cultivated. The mainstream narrative often prioritizes experienced designers with proven AI proficiency, viewing them as immediate assets in an AI-driven environment. This view, however, overlooks the crucial role of cultivating juniors, as described in the context of the nervous new hire learning critiques. Experienced designers are often adept at navigating existing systems, but the challenges emerging from AI’s variance problem and the loss of “screens” as a unit of measurement demand a different kind of problem-solving. True innovation, particularly in interaction patterns for unpredictable systems, often comes from fresh perspectives unburdened by established assumptions. The Oliver North home page, a first web credit in 1995, represents a time when foundational patterns were being invented, and designers were hired for their potential to learn rather than their pre-existing expertise. The industry’s current focus on immediate AI readiness in senior roles risks stifling the very adaptability needed to solve the next generation of interaction problems. Within two years, organizations that heavily invest in structured junior designer mentorship programs, specifically focused on qualitative user research for AI product variance, will demonstrate measurably higher user retention rates for their AI-powered features.

The primary opposing force to these evolving design requirements is the persistence of traditional, quantitative performance metrics. Executives and product managers, accustomed to measuring success by screen-based completion rates or output volume, resist the investment in qualitative research necessary to understand the nuances of AI variance or the long-term value of cultivating junior judgment. This resistance often manifests as pressure to deliver features quickly, prioritizing immediate, measurable efficiency gains over the harder-to-quantify benefits of trust and emergent design thinking.

A UI UX professional should integrate “variance narratives” into their user research process for any AI-powered feature. Instead of just observing task completion, ask users to describe their expectations of performance consistency and how they interpret deviations in speed or output. Document these narratives and feed them directly into requirements for AI system explainability and design system guidelines for communicating non-deterministic behaviors.

---

Curated References