KPMG Thought Leadership Assistant: AI-Powered Research for Client-Facing Professionals

Overview

Led end-to-end design of TLKA, an AI-powered research tool streamlining how KPMG professionals discover and leverage firm thought leadership. The platform enables rapid access to insights, client stories, services, and alliances, consolidating content discovery into a single AI-powered interface.

KPMG professionals were spending significant time on manual research across multiple systems before client engagements. TLKA replaced that fragmented process with a unified AI search experience that gets to the right content quickly and reliably.

The Problem

Senior managers and advisory professionals needed to locate relevant KPMG thought leadership, case studies, and service information rapidly, but the existing approach spread content across multiple disconnected systems. Time-intensive searches frustrated users, and professionals couldn't trust platform accuracy without cross-referencing multiple sources.

The result: knowledge workers were spending preparation time hunting for materials instead of actually preparing for client meetings, and comprehensive firm thought leadership was going underutilized.

Research

Conducted 12 interviews with advisory professionals to understand navigation challenges and information-seeking behaviors. Analyzed behavioral data and search patterns to identify friction points. Facilitated feedback sessions with client-facing teams to understand what "good" looked like in their specific workflows.

Key insight: users didn't want more filters or advanced search options. They wanted fewer steps and more targeted information, with transparency about where content was coming from so they could trust it without verification.

Design Solution

Designed a unified AI search experience delivering quick access to thought leadership, client stories, and service documentation. Introduced a streamlined interface with suggested prompts to guide users toward relevant content without requiring precise queries.

Integrated source files and direct links throughout results to build credibility and allow users to verify and share content directly. The interface was designed to feel like a knowledgeable colleague rather than a search engine.

Testing and Iteration

Conducted usability testing with 30 professionals across Advisory and Tax practices. Ran A/B comparison testing on search result layouts and source attribution placements. Collected ongoing feedback via WalkMe and in-tool surveys, iterating on prompt suggestions and result hierarchy based on real usage patterns.

Outcome

Reduced time searching for content, with users reporting significantly faster research-to-preparation workflows. Increased user satisfaction with research tools. 60% adoption rate among client-facing professionals within three months of launch.

Key Learnings

Simplification drives efficiency. Reducing steps and delivering targeted information matters more than adding search controls.

Trust requires transparency. Source attribution and content tagging are not optional extras; they are core to whether professionals will act on AI-generated results.

Continuous feedback ensures relevance. Regular in-product feedback loops kept the platform aligned with how professionals' needs evolved over time.