Erica Hellbe[email protected]

Pilot · Founding designer · B2B AI

Making noisy product signals feel actionable.

Pilot watches analytics, customer sentiment and shipped code to surface product opportunities while keeping the evidence close enough to earn a team’s trust.

The design problem

Make AI useful without asking teams to trust a black box.

The challenge was not displaying more information. It was helping a team understand what deserves attention, why it matters and what they could do next.

Product principle
Lead with the takeaway while preserving the evidence behind it.
Control
AI proposes; product teams decide what deserves action.
Pilot suggestion detail showing an event-sharing problem, annotated interface and ticket creation controls
A suggestion keeps the product problem, interface context and next action together.
Pilot Product Tracker showing shipped changes, search coverage trends and metrics by product area
Product Tracker connects shipped changes to KPI movement across the product.

Designing for trust

01

Compress, don’t conceal

Lead with the takeaway while preserving the evidence behind it.

02

Make hierarchy do the explaining

Separate urgent signals, recurring themes and ambient context.

03

Keep people in control

AI proposes; product teams decide what deserves action.

Next projectZeal AI