Unified UX: Designing a Single Experience for an AI Support Suite

Unifying AI Escalation, Case Quality, Agent Assistance, and Knowledge into One Coherent System

Overview

As AI capabilities expanded across the customer support lifecycle, multiple intelligent agents were introduced to solve distinct problems—real-time agent assistance, escalation prediction, case quality auditing, and knowledge discovery.

While each capability delivered value independently, the overall experience began to fragment. Users were interacting with multiple AI tools that behaved differently, required separate learning, and introduced operational and security overhead.

This case study focuses on how I led the design of a unified UX instance, bringing together:

AI Agent Partner (Agent Helper)
AI Escalation Manager
AI Case Quality Auditor
AI Knowledge Agent

into one coherent, scalable experience.

The Core Problem

The problem was not visual inconsistency—it was systemic fragmentation.

As multiple AI capabilities coexisted, the platform suffered from:

Different mental models for similar AI actions
Inconsistent workflows for agents, admins, and supervisors
Repeated learning effort as new AI capabilities were introduced
Increased cognitive load during live case handling
Inconsistent transparency into AI decisions, affecting trust
Longer onboarding and demo cycles

Beyond UX, this fragmentation also created business and operational risks:

Each AI capability required a separate Salesforce package upload
Salesforce security reviews are costly and time-consuming
Customers were hesitant to install multiple packages due to security concerns
Release velocity slowed due to repeated compliance checks

At scale, the platform risked becoming powerful but expensive, complex, and hard to trust.

My Role

I led both platform-level experience strategy and feature-level execution for the unified AI instance, operating as a Product Manager and UX Lead across multiple AI agents.

At any given time, I was responsible for 5 AI agents spanning 3 core product areas, which required balancing long-term system coherence with immediate feature delivery.

My responsibilities included:

Defining a shared experience vision across AI Agent Partner, Escalation Manager, Case Quality Auditor, and Knowledge Agent
Designing and executing a unified interaction model for AI behavior across live cases, post-resolution workflows, and admin configuration
Owning feature-level UX and product decisions, from problem framing to final delivery
Establishing experience principles to maintain consistency while shipping at speed
Collaborating deeply with product, engineering, security, QA, and GTM teams
Governing craft quality while actively designing, reviewing, and iterating on features

This work required me to constantly switch between:

Platform thinking (coherence, scalability, cost, security)
Feature execution (shipping real workflows, edge cases, and states)

In short, I wasn’t defining the system from a distance —
I was inside it, building and scaling it at the same time.

Reframing the problem

I reframed the challenge from:

“Each AI capability needs a better UX”

to:

“The AI suite must behave like one intelligent system, not four separate tools.”

This shift unlocked alignment across teams and allowed us to think in terms of shared behavior, not isolated features.