UseBelha: making AI-assisted customer operations easier to follow.

Customer-operations tools often force people to switch between conversations, tasks, status screens, and disconnected systems. AI adds another layer of complexity when its actions are not visible or easy to review.

I explored how the product could organise information and communicate what the agent is doing, what needs attention, and what the operator can do next.

The product direction was evolving, and some customer and implementation details were not public. I therefore kept this portfolio story at concept level and avoided claiming confidential outcomes or invented performance metrics.

Prioritise actions over volume

The interface direction emphasises current status, exceptions, and next steps rather than treating every event as equally important.

Keep the agent observable

Progress and activity should be understandable so operators can build trust and know when human judgment is required.

Use plain language

Labels and supporting copy should explain customer impact, not expose internal technical terminology.

The work produced an initial interface direction and a clearer hierarchy for discussing the customer-operations experience. It remains concept work; no live adoption metric is claimed.

A stronger next version would test the concepts with operators using real scenarios, then replace assumptions with observed task patterns and clearer success measures.