The Invisible Customer: UX Design for Algorithms
Exploring how the rise of AI agents is reshaping design beyond human aesthetics.
May 21, 2026
May 21, 2026
•4 mins read
Designers traditionally visualize a human on the other side of the screen, scrolling, making choices, interacting. It was all about crafting a seamless, delightful experience to guide their decisions. However, there's a shift happening: increasingly, the customer isn’t always human.
Recently, I reflected on how design work feels somewhat empty, as if something essential is missing. I initially thought it was a lack of humanity. While that remains partly true, there's another angle: the real audience might not even be a person anymore.
Now, digital agents, automated software helpers, often make decisions before humans do. AI 'buyers,' browser bots, and filter engines are sorting and selecting before a human interacts with a product. The customer is often invisible.
Designing for something without an appreciation for beauty or storytelling presents unique challenges. An AI searching for a table doesn’t care about warmth or style; it’s about matching parameters. An agent's 'experience' is an API call, not an emotional journey. Elements like the perfect shade of blue or clever copy only matter if they align with a rule the agent recognizes.
To test this, I experimented with OpenClaw, an agent engine, over a weekend. Could it learn my taste for 'warm minimalist' furniture? The process was strikingly different from human decision-making. The agent ignored images with showroom lighting, penalized grid-like layouts, and favored descriptors like 'natural aging of oak.' Eventually, it discovered a small Danish furniture maker I hadn’t known.
The agent’s logic was driven by parameters, not feelings. It linked my 'taste' to criteria it could quantify, lighting, text, provenance, without understanding why those mattered to me.
Design for agents shifts the focus. It's no longer about guiding a human through emotional highs but creating instructions for a machine detective. For humans, stories matter deeply; for agents, metadata is key.
Traditionally, design and branding assumed an emotionally-engaged audience. Yet, if agents increasingly make initial choices, the web transitions from a gallery of experiences to a data landscape. Stories are reduced to tags, and brands to parsable data.
While some may argue that humans remain at the decision's end, it’s different. Agents filtering the initial choices strip away subtle, nonverbal signals. Your brand morphs into character sets and APIs, not rich narratives.
This isn’t entirely negative. Efficient agents reduce the time spent navigating manipulative sites. Conversely, bad agents could make the web more homogeneous, optimized for machine-friendly profiles instead of diverse human experiences.
The surprising change is not just humans losing interest but their roles being automated. In few years, perhaps half of your 'customers' are actually bots representing the people you aim to reach.
To excel in this era, recognize that the audience is splitting: one part machine, one part human. Both are crucial, but one remains unseen.