The memory problem hiding inside “intent is the new interface”

A designer I respect said something to me a while back that stuck:

“Most products don’t have a UX problem. They have a memory problem.”

I didn’t buy it at first.

It sounded too simple. Like one of those lines that feels smart but doesn’t hold up once you get into the work.

I thought—no, we have flow issues. Too many steps. Not enough clarity. Bad handoffs.

All the usual things.

But the more I’ve worked on systems that are supposed to feel intelligent, the more I’ve realized…

we’ve been fixing the wrong layer.

We keep talking about intent

“Intent is the new interface.”

You’ve probably seen it. Heard it. Maybe even said it.

And it’s not wrong.

But it’s incomplete.

Because intent doesn’t just show up.

It forms.

Intent is what happens before the system ever sees you

Think about the last time you tried to do something that actually mattered.

Not a quick task. Something with a little weight to it.

You didn’t just open an app and decide.

You probably:

  • searched a few things

  • clicked into something and backed out

  • paused

  • compared options

  • changed direction halfway through

That’s the real moment.

That messy, half-formed thinking.

That’s intent.

And then we erase it

You finally engage.

You open the app. Walk into the branch. start the process.

And what happens?

You get asked:

“What are you here for?”
“Can you tell me a little more about what you’re looking to do?”
“Have you done this before?”

You already answered those questions.

Just not in a form the system recognizes.

So you repeat yourself.

Reconstruct your thinking.

Try to explain something that took you hours (or days) to figure out…

in a few sentences.

That’s the break.

Not in the interface.

In the continuity.

This is where AI exposed us

Over the last year, we’ve rushed to make everything “smarter.”

Faster answers. Better recommendations. More automation.

But here’s what actually happened:

We introduced intelligence into systems that don’t remember anything.

So now you get responses that sound right…

but don’t feel right.

Because they’re missing the one thing that matters:

context.

AI doesn’t fix this problem

It amplifies it

Without context, AI is just guessing faster.

It’s confident. It’s fluent. It’s efficient.

And it’s wrong in all the subtle ways that matter.

With context, something shifts.

It starts to feel like intuition.

Like the system was actually there with you.

We’ve been calling it intent

But it’s actually memory

Intent is the signal.

Memory is what makes the signal usable.

If a system can’t remember:

  • what you explored

  • what you ignored

  • what you almost did

  • what you cared about but never explicitly said

…it doesn’t understand you.

It just reacts to you.

When a system remembers, you feel it immediately

You don’t have to repeat yourself.

You don’t have to reconstruct your thinking.

You don’t feel like you’re jumping between disconnected moments.

It feels like something is carrying the thread with you.

That’s the bar now.

Not faster.

Not cleaner.

Not even “smarter.”

Continuous

This is where design actually changes

Not in adding AI features.

Not in better prompts.

Not in cleaner screens.

In deciding:

What should this system remember?
What actually matters vs what’s just noise?
When should context show up?
When should it stay invisible?

This is harder work.

It forces you into the messy parts:

  • fragmented systems

  • incomplete data

  • conflicting signals

  • real trade-offs

Which is exactly why most teams avoid it.

One belief I keep coming back to

AI is a waste of time without context.

And context is useless without memory.

Most systems today have neither.

So where this goes

We can keep designing better interactions.

Or we can start designing systems that don’t forget.

Because the moment a system actually remembers you

is the moment it stops feeling like software

and starts feeling like it understands.

Next
Next

The first launch of our Branch Transformation