AI Is Killing Design (And Designers Did It To Themselves)

AI didn't commoditize design. Designers did.

I gave a talk at DesignTO in January 2020 called "Our Future in the Intangible." The argument was that design has always been ripe for commoditization, and that technology was just accelerating a process already well underway. AI at that point was generating weird-looking cats on Google. We were nowhere near what we have now.

I recently resurfaced that talk on our podcast, from memory, five years later. What surprised me wasn't how much had changed. It was how little had. The structural forces I was describing in 2020 have simply become more visible, more urgent, and harder to rationalize away.

So let me make the argument plainly: we did this to ourselves.

How design expanded and then got squeezed

When I started at OCAD in 2006, the design domain was legible. Editorial, branding, motion, wayfinding. Digital was called interaction design. UX wasn't really a term yet.

Over the next decade, the domain kept expanding. Industrial design. Human factors. Service design. HCI. Data visualization. Cognitive science. Business modeling. Strategy. Design didn't just absorb adjacent disciplines, it became a container for almost anything involving human judgment and creative problem solving—the same breadth that underpins our focus on holistic user experience and long-term client relationships.

That expansion was genuinely exciting. IDEO was building methodologies to get business people speaking design and designers speaking business. Design thinking moved from being niche to being taught in MBA programs. Accenture acquired Fjord. Deloitte acquired Doblin. The signal was clear: design was being pulled upstream, toward strategy and away from execution.

Then the tools caught up.

Figma changed how design teams work more fundamentally than most people acknowledged at the time. It operationalized the design process, made it sequential, workflow-oriented, closer to how development has always run. The classic design tropes of exploration and open-ended iteration started fading. In their place: design systems, component libraries, repeatable tokens, documented patterns—the shift toward "design as process" that recent signals and summaries from Config make explicit.

Which are great. Until they become the default.

The assembly line problem

I keep coming back to Ford and Toyota when I think about where design agencies landed.

Ford democratized access to a vehicle by creating the assembly line. Toyota took that further and optimized every detail of the manufacturing process until the process itself became the differentiator. You don't buy a Toyota because of a standout feature. You buy it because Toyota has earned a reputation built on operational consistency.

The analogy is uncomfortable for agencies because we've done something similar. Design systems brought us frictionless production. Reusable components reduced time to delivery. Standardized UI patterns lowered the skill floor. All of that is objectively useful. But in optimizing for repeatability, we stripped out a lot of what made the output distinctive. And when you've made design repeatable and scalable, you've made it automatable.

That's where AI entered. We handed it a well-organized kitchen and were surprised when it started cooking.

What design commoditization actually means

Design commoditization is the process by which design services become standardized and undifferentiated enough that price becomes the primary selection criteria. When clients can't tell the difference between your custom solution and a template from a UI kit, that's commoditization at work.

This happened across three channels at once, and it reshaped how organizations approach digital design services, from websites to complex platforms.

Tools became accessible to non-designers. Squarespace, Webflow, Framer, generative AI interfaces. Any of these can produce something that looks professional to a client who doesn't know what professional used to require.

The gig economy made design talent fungible. Upwork and similar platforms turned skilled execution into a global price comparison. If you're billing time and materials, and another designer somewhere else can do the same task faster and cheaper, the math doesn't work in your favor.

Nearshoring brought development costs down dramatically. By 2026, what I was describing as an emerging trend in 2020, Latin America as a serious technology production hub, is fully realized. Columbia, Mexico City, Peru: established centers for development agencies, remote teams, acquired firms, all competing with more established development services and full-stack digital delivery teams. That drove the cost of delivery down further, and it particularly affected agencies whose value proposition was primarily execution.

The result is a split. Low-paid production work and high-value strategic work, with less and less in the middle, a pattern that's consistent with broader shifts in the future of the design industry.

Where value actually lives now

Here's what I don't think AI replaces: the ability to define what's worth building and why.

Execution has a floor, and AI is getting very close to it. But the design process at its best has never been primarily about execution. It's been about user research, defining the problem statement clearly, understanding what the audience actually needs versus what they say they need, and then making judgment calls that no amount of data analysis can make for you.

The designers and agencies who are genuinely insulated from commoditization are the ones who've been operating at that level all along. They're not competing on output. They're competing on diagnosis.

This is where the discipline of market leaders framework becomes useful. I've been drawing on this for years, from the book of the same name by Treacy and Wiersema. It argues that market leaders generally win through one of three disciplines: operational excellence, product leadership, or customer intimacy. Amazon, Google, Accenture are the archetypes. The key insight is that you have to choose one and build everything around it. Trying to lead through all three is how you become mediocre at everything.

For design agencies, this is now existential, not aspirational. If your value proposition is operational excellence, cheap and fast, you are competing with AI directly and will lose. If it's product leadership, you need to be genuinely building better things and commercializing faster than the field. If it's customer intimacy, you need relationships deep enough and services broad enough that the client isn't shopping around on price.

At Tennis, we've landed firmly in customer intimacy territory for our B2B web design and product development work. That means going deep with a narrower set of clients, running proper discovery, building governance frameworks, and staying engaged post-launch. It's not the only viable model, but it's the one that scales with trust rather than volume.

The network effect and the lean team

One thing the commoditization shift has produced that I find genuinely interesting: the viability of very small, very focused teams.

In 2020, the gig economy was enabling businesses to loosen the coupling between their core operations and their specialist service providers. Tightly coupled vendor relationships were giving way to loosely coupled ones, and some things were going fully uncoupled into platforms like Upwork. At the time, this was a cost-cutting pattern. Now it's a design principle.

Look at Lovable. Around 30 people, valuation in the billions, building AI-powered product design tools. That's not an anomaly. The network effect of AI, being able to compress what would have taken a team of twenty into what a focused team of five can now do, is producing leaner and leaner high-output organizations. This is going to keep happening.

The agencies that understand this are building networks of loosely coupled specialists rather than trying to grow headcount, often leaning on focused partners for strategy, UX, and digital transformation services. You keep what differentiates you internal. Everything else, you find the best node in the network and plug it in.

That means being ruthlessly honest about what your core actually is.

So what do you do with this?

The worst response to commoditization is to add more services. More capabilities on the pitch deck, more verticals in the case studies. That's the instinct, and it's wrong. Variety without focus is just dilution.

Niching down is uncomfortable because it feels like leaving money on the table. We did it at Tennis, shifting from "digital agency" to "B2B web design and product development agency" with a focus on website design and development for marketing teams. Still a broad niche. But it clarified who we serve and how we differentiate, which makes every other decision easier.

The question for any designer or agency right now is the same one I was asking in 2020, just with more urgency behind it: what do you do that artificial intelligence cannot do more cheaply?

If the answer is "execute familiar patterns quickly," that's a problem. If the answer is "understand what clients actually need, align stakeholders around it, build systems that outlast the engagement, and operate as a strategic partner long after launch" that's a position worth defending, and it depends on a rigorous web development process that treats sites as long-term products.

The value of design has always been in the thinking, not the making. Commoditization just finally made that obvious, which is why ongoing articles, podcasts, and resources on digital strategy and UX matter more than ever.