The Scam of Seamlessness

You’ve signed up for a service. The onboarding took ninety seconds. A welcome email arrived before you found the tab. Every form field autocompleted. Every transition animated. The loading indicator was so brief you almost missed it. No rough edges. No awkward pauses. No moment where a human being had to touch your account. It felt efficient. It felt modern. And somewhere in the back of your head, a quiet alarm went off.

That alarm is not a bug. It’s the new baseline response to frictionless digital experiences. And if you’ve built one, it’s worth asking what signal it’s sending.

The decades we spent removing all the bumps

For most of the last fifteen years, the goal of UX design was clear: reduce friction. Remove steps. Speed up load times. Autofill the forms. Make the interface so smooth that users get from intent to outcome without noticing they moved.

This was the right goal for the era. When the internet was slow and clunky and every extra click meant drop-off, friction was a real cost. Removing it was a genuine service. Amazon’s one-click patent. Stripe’s seven-line checkout. These were not design tricks… they were trust builders, because they demonstrated care for the user’s time.

The gospel of seamlessness spread. By 2023, every startup pitch included a demo with zero loading states and a user journey that felt like sliding on ice. Frictionless became a proxy for professional. For premium. For competent.

Then AI happened. And the proxy flipped.

What seamlessness signals now

AI generates frictionless outputs by design.

Automated content arrives instantly. Automated responses are smooth and complete. Automated onboarding flows have no rough edges because no human being made them on a particular Tuesday with competing priorities and a slightly misread brief. The absence of friction is, in the most literal sense, what automated production looks like.

This does not mean users have consciously decided that your clean interface is suspicious. Most of them haven’t. What it means is that their evaluation of your experience has shifted... the same way it shifted for headshots. They’re not running detection. They’re running a different question: is anyone actually here? Frictionless used to answer yes. Now it answers nothing.

The food blog Inspired Taste learned this the hard way from the wrong direction. After Google AI Overviews launched and began synthesizing recipe content without attribution, the blog saw an 80% traffic collapse and an 80% revenue collapse — nearly simultaneously. Google’s AI was serving frictionless recipe summaries: clean, complete, instantly available, no rough edges. Users got what they needed. Inspired Taste, which had tested those recipes against actual chemistry in an actual kitchen over years of publishing, became invisible.

The product that replaced them was smoother. It was also a machine averaging training data. And for a while, nobody noticed the difference.

What friction actually does

The knitting community figured this out faster than most.

When AI-generated patterns started flooding Ravelry and Etsy, the community didn’t respond with better detection tools. They responded with a standard: video documentation of the physical creation process. Hands on yarn. Real needles. Real mistakes mid-row. They called it “Proof of Stitch.”

The logic was simple: AI generates patterns that look plausible on the page but cannot be physically executed — the seams don’t meet, the topology is impossible. A human who knitted the pattern had to encounter reality. The friction of physical constraint is proof that a real person did the work.

This is what friction does that seamlessness can’t: it leaves evidence of contact with the actual world.

A slightly slower response time because a real person reviewed your request. A confirmation email that uses your specific details in a non-templated way. A welcome message that arrives the next morning with something only possible if someone read what you wrote. A checkout process with one step that requires a real decision rather than a default.

None of these are bad UX. They’re evidence that a human being is handling the request.

The split is already happening

The internet is bifurcating, and the split is not about quality. It’s about friction.

On one side: the Slop Web. Free, instant, algorithmically distributed, frictionless, and trusted approximately as much as a banner ad. Volume is infinite. Marginal cost is zero. Every experience is smooth because no human being had to touch it.

On the other: a smaller, gated, often slower, clearly human-operated web where friction is the signal. Substack newsletters that arrive on irregular schedules because the writer had a harder week. Response times that suggest someone is thinking, not auto-responding. Onboarding calls that are actually calls. Waitlists that exist because capacity is real.

The premium signal used to be polish. It is increasingly becoming proof of effort. Not fake effort, not performed effort: actual evidence that someone was there.

Where to put a speed bump

I’m not arguing for bad UX. Unnecessary friction, the kind that exists because no one tested the flow, or because a developer copy-pasted a form from 2012 — is still a cost worth eliminating.

The question is whether there’s a moment in your customer journey where a deliberate manual touchpoint proves that a human is on the other side of the transaction. Not to slow things down. To leave evidence.

The brands that are navigating this well have made peace with a counterintuitive idea: making something slightly harder to access makes it harder to dismiss as automated.

What does that look like in your product?