Enabling Scalable Innovation: UX for AT&T’s Multi-Agent GenAI Platform

AT&T sought to streamline operations and foster innovation through a multi-agent generative AI platform. As UX Lead, I spearheaded the end-to-end design and strategy to make this complex system intuitive, inclusive, and scalable. By tackling key challenges in accessibility, discoverability, and workflow efficiency, my work helped reduce technical barriers, improve user engagement, and deliver a more cohesive enterprise experience.

 

Problem

Single-agent AI systems were too limited to address AT&T's complex, enterprise-wide challenges. Teams needed to automate workflows, enhance support, and accelerate development. However, early platform versions required deep technical knowledge, causing accessibility and adoption issues. Additionally, there were inefficiencies in modifying skills during project execution and limited component reuse, which slowed development cycles.

 

Solution

We designed a modular, scalable multi-agent platform with distinct, teachable AI agents orchestrated by a Group Manager. Each agent was specialized (e.g., coding, testing, support) and operated within a microservice architecture.

Key UX features included:

  • A shared library of reusable skills, agents, and teams

  • A robust search and permissions system for discoverability and collaboration

  • Pre-built starter kits to lower entry barriers for non-technical users

  • A split-screen execution interface for real-time skill editing

  • Drill-down chip-based navigation for intuitive hierarchy traversal

 
 

My Role

As UX Lead, I owned the end-to-end UX strategy from concept to launch. I collaborated with Product, Engineering, and Leadership to:

  • Lead user research and contextual inquiries

  • Create and test wireframes, flows, and interactive prototypes

  • Develop a comprehensive design system

  • Facilitate cross-functional alignment through workshops and agile ceremonies

  • Establish UX metrics to track adoption, satisfaction, and engagement

 

Key Responsibilities

User Research & Advocacy

  • Conducted research to identify user needs and pain points

  • Facilitated design thinking sessions to inform direction and prioritization

Wireframes & Prototypes

  • Built high-fidelity interactive prototypes using Figma

  • Supported quick iteration and stakeholder feedback loops

Content & Service Design

  • Mapped service blueprints

  • Wrote UI copy, instructions, and error messaging

Agile UX Leadership

  • Participated in PI planning, stand-ups, and backlog grooming

  • Delivered incremental design solutions aligned with sprint goals

Design System Creation

  • Leveraging other Figma design systems, I built and maintained a centralized component library for consistency

Accessibility & QA

  • Ran accessibility audits, QA reviews, and usability testing

  • Drove continuous UX improvements post-launch

 

CHALLENGE #1

Lowering the Barrier to Entry

Problem: Early users had to code skills from scratch using Python, limiting accessibility and delaying adoption.

Solution: Introduced a shared library and starter kits with ready-to-use agents and skills. Users could search, request access, and build on existing components.

Outcome: Reduced learning curve and setup time. Empowered non-technical users and promoted system-wide collaboration.

 

CHALLENGE #2

Improving Iteration During Execution

Problem: Users had to leave the execution screen to edit skills or agents, disrupting flow and reducing iterative refinements.

Solution: Designed a split-screen layout for simultaneous execution and editing, plus a drill-down navigation to simplify structure and access.

Outcome: Streamlined workflows and reduced cognitive burden. Enabled users to refine outputs in real time.

 

Impact & Outcomes

The impact was observed qualitatively through improved usability testing feedback, increased adoption in demos, and broader engagement across technical and non-technical teams.

  • Improved satisfaction and engagement through inclusive design

  • Faster onboarding and project setup with starter kits

  • Enhanced usability and performance via real-time editing tools

 

Final Reflection

This case highlights how UX can transform advanced AI systems into usable, scalable enterprise tools. By embedding myself across research, strategy, and execution, I helped AT&T build a future-ready platform that prioritizes users as much as innovation.

Case Studies