A knowledge assistant for financial advisers
The problem
Decades of research and internal expertise sit across tens of thousands of documents; advisers couldn’t retrieve the right answer fast enough during client work.
The AI approach
A GPT-4-based assistant performs retrieval over the firm’s curated research and knowledge library, returning sourced answers in natural language. A companion tool (“Debrief”, 2024) summarises client meetings and drafts notes and follow-ups, with client consent.
Evidence it works
Launched September 2023 for 16,000+ advisers over a library of 100,000+ proprietary documents, the assistant reached over 98% adoption among advisor teams (per OpenAI’s own case study), with document-retrieval efficiency rising from roughly 20% to 80% and answers returning in seconds rather than minutes.
What “good” looks like
Faster, accurate retrieval grounded in approved internal content with citations; advisers spending less time searching and more with clients.
Feasibility & cost shape
Depends on a clean, well-governed internal knowledge base — usually the real project. Integration and compliance review are significant.
The clearest enterprise template for “RAG over your own knowledge.” We’d spend most of the effort on the knowledge foundation and governance, not the model.
Based on publicly reported information about the Morgan Stanley (with OpenAI) work.
This is an industry example included for illustration. It is not a Leia Intelligence project, and no client of ours is implied. Figures are as publicly reported by the original parties.
Sources: Morgan Stanley (newsroom) · OpenAI (customer story)