73% of Designers Say AI Collaboration Will Be the Biggest Impact in 2026

[maya_builds]

I just came back from a design leadership roundtable, and the sentiment was overwhelming: AI collaboration is the defining change for design in 2026.

Not AI replacing designers. Not AI generating final designs. AI as a collaborator - someone (something?) that helps us work differently.

The Numbers That Caught My Attention

A recent industry survey found:

• 73% of designers say AI as a design collaborator will have the most impact in 2026
• 93% are already using generative AI tools like ChatGPT and Midjourney in their current work
• 58% point to AI design assistants (like Figma AI) as the specific technology that will change how they work

But here’s the tension: 78% believe AI boosts efficiency, yet fewer than half feel it makes them better at their jobs.

Faster ≠ Better. That distinction matters.

What “AI Collaboration” Actually Looks Like

In my day-to-day, AI collaboration means:

The Tedious Stuff (Where AI Excels)

1. Layer organization: Figma AI can contextually rename and organize layers with a click. This used to take hours during handoff prep.

2. Content generation: Generating realistic placeholder text that matches the tone we’re going for, instead of lorem ipsum.

3. Accessibility checks: AI tools that catch contrast issues, missing alt text, and focus order problems automatically.

4. Layout exploration: “Give me 5 variations of this hero section” - not to use directly, but to spark ideas.

The Creative Stuff (Where AI Is a Starting Point)

1. First drafts: AI can generate a wireframe from a prompt. I never use it as-is, but it’s a faster starting point than a blank canvas.

2. Color exploration: “Suggest color palettes that work for a healthcare app” - again, starting points, not final answers.

3. Pattern recognition: AI can analyze our design system and suggest which components to use for a new feature.

The “Partner vs. Replacement” Question

At the roundtable, the most heated discussion was about where the line is.

I believe AI collaboration works when:

• AI handles exploration, humans handle selection - Generate 10 options, I pick and refine the best one
• AI handles consistency, humans handle innovation - AI ensures we follow the design system, I decide when to break it
• AI handles speed, humans handle judgment - AI gets us to 80% fast, I spend time on the 20% that matters

It breaks when:

• AI is expected to make taste decisions
• AI output goes directly to users without human review
• AI is used to replace design headcount rather than amplify design impact

The Efficiency vs. Quality Tension

The stat that fewer than half of designers feel AI makes them better at their jobs is telling.

My interpretation: AI accelerates production, but production was never the bottleneck. Understanding users, making strategic decisions, building consensus - these take the same amount of time with or without AI.

If your design process was:

• 20% understanding the problem
• 30% exploring solutions
• 50% producing artifacts

AI might compress that 50% to 15%. Great efficiency gain.

But if your design process was:

• 50% understanding the problem
• 40% aligning stakeholders
• 10% producing artifacts

AI barely moves the needle.

What I’m Watching in 2026

1. Agentic AI for users: 60% of designers think AI agents that act on behalf of users will have major impact. This changes what we design - we’re designing for AI intermediaries, not just human eyes.

2. Figma Make: The promise of keeping design and code connected, with AI bridging the gap. If it works, it fundamentally changes the design-to-development handoff.

3. Explainable AI: People won’t trust systems they can’t understand. As designers, we’ll need to design for transparency - showing users why AI made certain decisions.

Questions for the Community

1. How has AI changed your design workflow in practice? Not the hype - the reality of day-to-day use.

2. Where do you draw the “AI vs. human” line? What decisions should never be delegated to AI?

3. Has AI made you better at design, or just faster? I’m genuinely curious if anyone has experienced the former.

The 73% number tells me the industry is embracing AI collaboration. The less-than-50% “better at their jobs” number tells me we’re still figuring out how.

[product_david]

Maya, this is a perspective I’ve been waiting to hear from the design side. As a product leader, I see the AI collaboration question slightly differently.

The “bottleneck was never production” insight is crucial.

In my experience, design velocity was rarely the limiting factor for shipping products. The bottleneck was always:

• Figuring out what to build (discovery)
• Aligning stakeholders on direction (consensus)
• Validating that what we built actually worked (research)

If AI makes production 3x faster but doesn’t touch those other areas, we’ve optimized the wrong thing.

But here’s where I see AI actually helping product work:

1. Faster prototyping for user research: AI-generated mockups let us test more concepts with users before committing. We ran 3x more concept tests last quarter because mockup creation wasn’t the bottleneck anymore.

2. Variant exploration: Instead of debating which direction to pursue, we can now mockup all of them quickly and let user data decide. Less opinion-based design decisions.

3. Async collaboration: When I describe what I’m imagining in a Slack message, AI can generate a rough visual. The designer then refines it. We skip the “let me schedule a meeting to explain what I’m thinking” step.

Where it’s not helping:

AI can’t sit in a customer interview and notice when someone hesitates before answering. It can’t feel the frustration in a user’s voice during testing. It can’t read the room when the CEO has concerns they haven’t articulated yet.

The human judgment part of design - understanding why something works or doesn’t - remains firmly human.

On the “faster vs. better” question:

I think AI has made our design team better at one specific thing: exploring more options before converging. Previously, we’d generate 2-3 concepts due to time constraints. Now we might generate 10 and then select.

More exploration before convergence often leads to better outcomes. But it’s not the AI that made us better - it’s the process change the AI enabled.

[alex_dev]

As an engineer who works closely with designers, I want to share how this AI collaboration looks from the implementation side.

What’s actually changing in design-to-dev handoff:

The promise of tools like Figma Make - where design and code stay connected - is genuinely exciting. But we’re not there yet in practice.

What I see today:

• AI generates design mockups quickly
• Designer refines them
• Handoff to engineering still involves the same “what did you actually mean here?” questions
• AI-generated designs sometimes have inconsistencies that are invisible in mockups but obvious in code

The consistency problem:

When designers manually create components, they internalize the design system constraints. They know why certain patterns exist.

When AI generates layouts, it can produce things that look right but break conventions in subtle ways. Spacing that varies unexpectedly. Color shades that are close but not quite from the palette. States that work in isolation but conflict with other patterns.

I’ve spent more time recently asking “is this intentional?” about AI-generated designs than I did with fully human-created ones.

Where AI collaboration is helping engineering:

1. Design tokens are more consistent: AI tools that enforce design system compliance mean fewer “this padding should be 16px, not 18px” conversations.

2. Responsive variants faster: Designers can now provide mobile, tablet, and desktop variants more quickly, which reduces guesswork on our end.

3. Accessibility built-in: When AI catches contrast issues before handoff, we don’t discover them during QA.

My concern about “designing for AI intermediaries”:

Maya mentioned designing for AI agents that act on behalf of users. This is going to require new mental models.

Today, we design for humans looking at interfaces. Soon, we may need to design for AI parsing interfaces on behalf of humans. That’s a fundamentally different design challenge.

Are design teams ready for that? I’m not sure engineering teams are, either.

[cto_michelle]

I want to zoom out and address the strategic implications of AI collaboration in design, because there’s a headcount question lurking beneath this discussion.

The uncomfortable executive conversation:

When 73% of designers say AI will have the biggest impact, some executives hear: “We can do more with fewer designers.” I’ve been in those conversations.

The reality is more nuanced. AI collaboration enables:

• Same team, more output: The efficiency gains Maya describes
• Smaller team, same output: The cost-cutting interpretation
• Same team, higher quality: The aspirational version

What actually happens depends entirely on how leadership frames the investment.

What I’ve seen work and fail:

Worked: A company used AI efficiency gains to shift their design team from production-heavy to research-heavy. Same headcount, but more time spent with users. Product quality improved measurably.

Failed: A company cut design headcount by 30% because “AI can do the production work.” Within 6 months, design quality declined, engineering rework increased, and they were hiring again.

The difference: the first company understood that AI creates capacity for higher-value work. The second thought AI replaces the work.

On the $54.93 billion UX services market:

That market is growing because companies are realizing UX matters, not shrinking because AI makes designers unnecessary. The market for design judgment and strategy is expanding even as the market for design production is being commoditized.

Designers who focus on production are at risk. Designers who focus on strategy, research, and judgment are in higher demand than ever.

My advice for design leaders:

Frame AI collaboration in terms of what your team can now do that they couldn’t before, not in terms of how many fewer people you need.

• “We can now test 5 concepts with users instead of 2”
• “We can now provide more responsive variants at handoff”
• “We can now spend more time on accessibility and less on production”

That framing protects your team and genuinely delivers more value.

To Maya’s question about AI making designers better:

I think it depends on the designer. AI is making skilled designers more productive. It’s not making average designers more skilled. The gap between great and average may actually be widening.