Infinitus · AI Healthcare · 2022–Present · Patent Co-Inventor

AI Copilot

Infinitus automates healthcare phone calls using AI — cutting days of manual work down to minutes. As Staff Product Designer and Design Manager, I led the design of the AI Copilot: the internal tool that gives operations teams visibility, control, and trust in an agentic system handling millions of calls.

My Role

Staff Product Designer & Design Manager

Team

VP of Product
Engineering leads
Operations & QA teams
ML & data science

Methods

Jobs-to-be-done research
Workflow & systems mapping
UX/UI/IXD
Design prototyping
Usability testing
Design QA

The Problem

AI at scale needs human oversight — and that oversight needs design

Infinitus's AI agent handles millions of healthcare phone calls on behalf of providers — verifying benefits, scheduling, and retrieving patient information. As the system scaled, operations teams needed a way to monitor, review, and intervene in AI-handled calls without slowing the throughput that made the product valuable.

The challenge: design a Copilot interface that surfaces the right information at the right time, builds operator trust in autonomous decisions, and enables fast, confident action when human judgment is needed.

Process

Understanding human-AI collaboration at work

Discovery
Workflow Mapping
Concept Design
Prototyping
Iteration

I began with deep discovery sessions with the operations and QA teams — the primary Copilot users. Their core need wasn't just visibility into call status; it was confidence. They needed to know when to trust the AI and when to step in, without reviewing every call manually.

Workflow mapping sessions revealed that the existing tooling forced operators to jump between multiple internal dashboards to piece together call context. This created lag in escalation decisions and introduced error risk in high-stakes healthcare workflows.

From there, I led concept design and rapid prototyping cycles with engineering and ML partners, pressure-testing assumptions about when AI confidence scores were meaningful vs. misleading to non-technical operators.

Solution

A Copilot that makes AI legible and actionable

The AI Copilot consolidates call monitoring, live transcription review, escalation queuing, and outcome auditing into a single, unified interface. Key design decisions included:

  • Confidence-aware UI — visual indicators that communicate model uncertainty in plain language, not raw percentages, so operators can triage quickly without ML expertise.
  • Contextual escalation flows — one-click intervention paths that preserve call context, reducing the time-to-escalate from minutes to seconds.
  • Audit trail design — structured call summaries that satisfy compliance requirements while remaining scannable for ops teams under volume pressure.

The patent I co-invented covers a novel method for routing AI-generated call outcomes through human review loops — a direct product of the design systems work on this project.