Support Automation Assistant

Created a guided assistant to reduce escalations and speed up resolutions.

Role

Lead UX Designer

Team

2 engineers, 1 content strategist

Tools

Figma, Notion, UserTesting

Constraints

Limited AI stack; needed deterministic workflows

Outcome

Improved first-contact resolution by 18%.

Problem

Support agents faced repetitive triage tasks and inconsistent escalation notes.

Automation had to work with existing macros and ticketing APIs.

Shadowed agents, mapped decision trees, and iterated on microcopy with support leads.

Built a structured assistant that surfaces next steps, templates, and escalation triggers.

Provided clear status messaging and reduced cognitive load with progressive cues.

Average handle time dropped by 15% with higher satisfaction scores.

Agent trust increases when the system explains why recommendations appear.