Your AI detected something.
Great.
But what happens next?
A lot of AI features stop at:
Detect → Insight → Recommendation
Then the user has to figure out the rest.
That’s where AI UX can do more.
Detection → Action
The Dyson CameraJet example in this episode illustrates the idea:
Camera → AI detects gap → Targeted jet
The system doesn’t simply identify the problem.
It acts on the detection.
That can remove several steps from the user’s mental workload.
Instead of:
Find → Understand → Decide → Act
the system can handle more of the loop.
Less interpretation.
Less effort.
Less friction.
But don’t automate everything
Automation isn’t automatically good UX.
For more consequential actions, users need:
Visibility — What is happening?
Control — Can I stop or override it?
Feedback — What happened?
If the product acts on someone’s behalf, the action shouldn’t disappear completely into the background.
Users should know what the system detected, what it did, what happened and what they can do next.
Use the AI UX model
SENSE → DECIDE → ACT → FEEDBACK → CONTROL
Ask:
Can the system detect the problem?
Can it make a reliable decision?
Can it take a useful, low-risk action?
Can the user understand what happened?
Can the user intervene when it matters?
And remember:
Don’t ask:
“Can AI do this?”
Ask:
“Should AI do this automatically?”
The goal isn’t maximum automation.
It’s useful automation.
Comments