Company
EY is one of the world's largest professional services firms, with hundreds of thousands of consultants generating insights across assurance, tax, consulting, and advisory services. As a result, decades of accumulated expertise sat dormant in slide decks and workpapers, inaccessible when needed most.
The Challenge
EY consultants spent an average of 15 minutes searching for relevant precedents—and often came up empty, recreating analyses that already existed elsewhere in the firm. The root causes were familiar: unstructured content scattered across SharePoint, inconsistent metadata, and no reliable way to surface insights from past projects. Early attempts with LLMs introduced new risks, particularly around hallucinations when handling sensitive client information. With research showing 47% of professionals waste 1-5 hours daily searching for information (Source: Pyron), EY risked continued productivity losses and missed opportunities to leverage its institutional knowledge for competitive advantage.
The Solution
Maheep Bhalla, a Principal Product Strategist and member of A.Team's network, brought expertise in enterprise GenAI implementations and building trust in AI systems. His experience navigating compliance requirements in professional services proved invaluable.
Bhalla designed a phased approach starting with high-value use cases and canonical document types rather than attempting to index all content at once. He implemented RAG with embedding search across a multi-tiered vector index, working closely with compliance teams on security and SMEs on prompt patterns. The rollout progressed from internal alpha to a 35-user pilot, with careful tracking of satisfaction metrics and hallucination risks. This methodical approach built organizational trust alongside technical capabilities, ensuring the system would actually be adopted.
Technologies used:
OpenAI GPT-4: Core language model for understanding complex queries and generating contextual responses
Azure OpenAI Service: Enterprise deployment ensuring data security and compliance with client confidentiality requirements
LangChain: Orchestration framework for building retrieval pipelines and managing prompt chains
Pinecone: Vector database for storing and querying document embeddings at scale
SharePoint & Azure Blob Storage: Source systems integration for accessing existing knowledge repositories
Figma: UX prototyping and user testing workflows for the interface design
"Surfacing deep institutional knowledge via GenAI required precision, trust, and tight collaboration across AI, compliance, and user teams. We weren't just deploying LLMs—we were rethinking how knowledge travels inside a global firm."
— Maheep Bhalla, Principal Product Strategist
The Results
Bhalla helped EY transform its knowledge management capabilities from a manual search process into an intelligent discovery system that accelerated consultant productivity and client value delivery.
Efficiency Gains
Reduced time-to-answer by 40% on common internal queries, cutting average search time from 15 minutes to under 90 seconds
Knowledge Reuse
Increased reuse of past deliverables by 25%, eliminating redundant slide development and reducing manual SME escalations
Business Impact
Faster access to high-quality precedents contributed to improved win rates on proposals and margin improvements on client work
Strategic Value
The GenAI interface became a flagship KM investment, supporting EY's AI-led transformation narrative and reducing reliance on shadow knowledge networks