The Death of "Best Practices" in Product Design

We’ve all heard the phrase.

"Just follow best practices."

They're in design books, startup pitch decks, UX courses, and product playbooks. They've guided how we've built digital products for years.

And for a long time, they genuinely worked.

But something has changed.
The more AI reshapes products, the less these “best practices” actually work.

And in some cases… they are quietly hurting products.

For years, good product design followed a familiar set of principles:

These ideas made sense because most software behaved predictably.

Users clicked a button, the system responded, and everyone knew what to expect.

AI changes that relationship.

Today's products don't just display information. They generate it. They make recommendations. They take actions. Sometimes they even make decisions on the user's behalf.

And that's where many of our old design assumptions begin to break.And suddenly… fixed “best practices” don’t fit dynamic systems anymore.

Take one of the most common design principles:

Reduce friction.

Normally, that's solid advice. But AI introduces a different challenge.

Sometimes removing friction also removes understanding.
Sometimes hiding complexity makes users trust the product less.
Sometimes automation feels less like convenience and more like losing control.

A product can feel incredibly smooth...

...while leaving users wondering:

"Wait... why did it do that?"

That moment matters.

Because people don't build trust from speed alone.

They build trust by understanding what's happening.

This feels like the biggest change happening in product design today.

We've spent years looking for universal rules. But AI products don't live in universal situations.

They're adaptive.
They're personalized.
They're probabilistic.

The same design decision that improves one experience can completely undermine another.

A highly automated workflow might delight an experienced user. The exact same workflow might overwhelm someone using the product for the first time.

Neither approach is objectively right.

They're right—or wrong—depending on the context.

That's a very different way of thinking about UX.

A lot of AI products make an amazing first impression.

The demos are polished.
The interactions feel magical.
The automation is impressive.

But after a few weeks, users quietly stop relying on them.
Not because the AI isn't capable.
Because they don't fully understand what it's doing.

Many teams are still designing AI products using rules that came from a different era:

But AI doesn't always need less interaction.
Sometimes it needs better interaction.

There's a big difference.

It’s UX judgment.

The ability to decide:

This is no longer a design system problem.

It’s a decision system problem.

The teams that stand out over the next few years won't simply be the ones following best practices.

They'll be the ones who know when to challenge them.

They'll know:

They'll design for trust, not just speed.

They'll adapt experiences to people instead of forcing people through predefined flows.

And they'll treat users as collaborators, not just steps in a funnel.

Maybe "best practices" aren't disappearing.

Because in AI-era products:
consistency is no longer the goal. comprehension is.

And that might be the biggest shift product design has seen in years.

If you are designing, building, or shipping products right now, what comes next will matter more than what worked before.

The patterns are already here. How AI products fail. How behavior is replacing interface. How the best founders build. How the next generation of products gets made.

I write about them every Tuesday. Be the first to know.

futureofproduct.com

Get it free in your inbox →