Activation / Trust / Mobile UX

Equal AI Onboarding Drop-offs

A silent wait made activation feel broken.

Product
Equal AI assistant setup
Role
Product design lead
Scope
Onboarding recovery, status communication, assistant activation flow
Equal AI Onboarding Drop-offs case study cover

The tension

Users were not abandoning the assistant. They were abandoning an invisible wait that looked broken.

What was happening

The product needed users to complete call-forwarding activation before the assistant could create value.

What changed the brief

Observed behavior pointed to a trust problem, not a value problem: without progress or confirmation, users closed the app and later discovered setup had never finished.

Success looked like

Success meant more in-session activations, fewer exits, and users understanding what happened when setup failed.

The constraint

Some recovery actions could only be triggered by users, and the flow had to stay lightweight inside a mobile setup moment.

68.7%onboarding completion

More users finished in-session.

-47%app exits

Fewer exits during setup.

550kDAU scale

Designed for Equal's live assistant footprint.

The product story

The interface, one decision at a time.

01 / Before

Silence looked exactly like failure.

A carousel filled the wait, but never answered the only question that mattered: is setup actually working?

Equal AI onboarding activating assistant before state
The old wait entertained, but did not reassure.Users saw content cards while call forwarding happened invisibly, so progress was impossible to judge.

02 / The move

Turn waiting into a short product story.

Visible status proves the app is moving; a lightweight video uses the same 30 seconds to explain the value ahead.

Equal AI onboarding improved activation video state
Progress now has a visible destination.A clear activation state and a lightweight onboarding video make the wait feel active without heavy information.

03 / Soft recovery

If setup fails, diagnose before asking for patience.

After automatic retries, the product identifies likely causes, shows the affected SIM, and gives actions only the user can take.

Equal AI onboarding activation failed recovery state
Recovery names the problem and the owner.Airplane mode, network reset, restart, retry, and support form a calm ladder instead of a dead end.

Onboarding video

Onboarding in motion

The setup moment becomes visible, paced, and recoverable.

The trust test

The happy path is only half the product.

Forwarding setup fails

The screen names the issue, gives the next action, and keeps support close.

Network or SIM blocks activation

Users get a retry route with plain-language causes instead of a dead end.

The wait feels broken

Progress, proof, and status messaging replace silence during activation.

The decisions

Three moves carried the story.

01

Make waiting active

Progress, retry, and success states made the verification step legible instead of leaving users to assume the product had stalled.

02

Use setup time to teach value

A lightweight video and status copy explained what the assistant would do next without overloading the screen.

03

Recover with specific user actions

Failure states explained the issue, identified possible SIM or network causes, and offered a calm path to retry or get help.

What I would test next

The redesign works because it respects uncertainty. It does not hide failure; it makes failure understandable and recoverable.