This week’s design discourse zeroes in on uncomfortable truths: AI tools can now produce interface designs in seconds, but catching their mistakes requires entirely new workflows. Meanwhile, the industry is reckoning with who actually belongs on a design team, what screens truly are as a medium, and why accessibility remains everyone’s blind spot. We’ve pulled the most actionable insights from five current discussions shaping how our team approaches web design projects right now.

Key Takeaways

• Hiring the “best” CV for a design team often produces homogeneous thinking and weaker outcomes.
• Flat design’s rejection of skeuomorphism left a conceptual vacuum about what screens actually are.
• Accessibility expertise isn’t a prerequisite — every designer can and should start learning now.
• A new Figma-to-Applitools integration catches visual discrepancies in AI-generated UI before production.
• AI-generated design assets raise context-dependent ethical questions that designers must navigate deliberately.

Stop Hiring for the “Best” CV — Hire for Team Composition

Design leadership has a recurring blind spot: stacking teams with top-tier individual contributors who share near-identical skill sets. As explored in a recent piece on why the best resume may be the wrong resume for a design team, experience and expertise alone don’t guarantee innovation. Cognitive diversity — different problem-solving approaches, industry backgrounds, and creative instincts — drives better design outcomes. We apply this thinking when building project squads for our clients: pairing a conversion-focused UX specialist with a brand-led visual designer consistently produces stronger results than doubling down on one profile.

Rethinking What a Screen Actually Is

Flat design correctly identified that a screen isn’t leather or brushed metal. But the movement then defaulted to treating the screen as nothing at all — a blank, featureless void. A thought-provoking analysis on what a screen is actually made of argues this assumption stripped digital design of meaningful materiality. For our team, this is a practical consideration. When we design interfaces, we’re increasingly reintroducing subtle texture, depth cues, and spatial relationships — not as decoration, but as functional affordances that help users parse information faster. The pendulum is swinging back toward intentional physicality in UI, and client sites benefit from it.

Accessibility Is a Starting Point, Not a Specialism

Most designers dodge accessibility because it feels like compliance paperwork. That’s exactly the problem. A candid piece arguing that nobody starts as an accessibility expert makes the case that waiting for a dedicated specialist means it never gets done. We’ve embedded basic accessibility checks — contrast ratios, keyboard navigation, semantic HTML — into every stage of our design and build process. It’s not a separate audit. It’s part of the job. Designers who treat it as optional are shipping exclusion by default.

Figma and Applitools Now Catch AI Coding Mistakes Automatically

AI coding agents can translate a design into working front-end code remarkably fast. They also introduce subtle visual regressions — a misaligned button here, incorrect spacing there — that slip past manual review. A new integration covered in detail at Web Designer Depot’s report on Figma and Applitools turns Figma files into visual baselines and flags pixel-level deviations automatically. We’re evaluating this for our QA pipeline. Any team using AI-assisted development without automated visual regression testing is essentially shipping unchecked code.

AI-Generated Design Assets Demand Contextual Ethics

A missing-dog poster made with AI looks polished. It also looks hollow. As examined in a nuanced discussion on the ethics of AI-generated design depending on context and intention, the appropriateness of AI output hinges entirely on situation. A quick social media graphic? Probably fine. A brand identity or campaign centrepiece? That needs human craft and judgement. We advise our clients along these lines daily: use AI to accelerate production tasks, but keep human designers in charge of anything that carries strategic or emotional weight.

The common thread across all five discussions is clear: tools are getting faster, but design quality still depends on human decisions — who you hire, how you define a screen, whether you prioritise inclusion, how you verify AI output, and when you choose craft over convenience. As we move through 2025, our team continues to treat these as non-negotiable standards rather than optional extras.

Frequently Asked Questions

How do web designers check AI-generated code for visual errors?

Tools like the new Applitools integration with Figma convert original designs into visual baselines and automatically flag pixel-level differences in AI-generated front-end code. This catches spacing, alignment, and colour discrepancies before they reach production.

Why does accessibility matter in website design even without legal requirements?

Accessible websites reach a wider audience and deliver better usability for every visitor, not just those with disabilities. Ignoring accessibility means shipping a site that actively excludes potential customers.

What is cognitive diversity in a web design team?

Cognitive diversity means assembling designers with different problem-solving styles, industry backgrounds, and creative approaches rather than hiring identical skill profiles. It leads to more innovative design solutions and reduces groupthink.

How do web designers decide when to use AI-generated assets?

Context determines appropriateness: AI works well for rapid production tasks like placeholder imagery or internal mockups. Brand-critical or emotionally significant design work still requires human craft and strategic intent.