Morgan Stanley

Morgan Stanley wealth advisors were spending significant time after client meetings manually creating notes, action plans, and follow-up communications — work often delegated to junior employees. This administrative burden detracted from time available for client engagement and prospecting. With roughly 1 million Zoom calls per year across the wealth management division, the cumulative cost was enormous.

Morgan Stanley built Debrief using OpenAI's GPT-4, an AI assistant that joins client Zoom meetings to transcribe discussions and automatically generate detailed meeting logs, draft follow-up emails, and conversation summaries. Client consent is required before each session. The tool required months of prompt engineering to fine-tune and was planned for rollout to the firm's approximately 15,000 advisors by July 2024.

Advisors in the pilot reported saving approximately 30 minutes of post-meeting work per session. With 1 million annual Zoom calls, the aggregate time reclaimed across the advisor force is substantial. Morgan Stanley expects the productivity gains to translate into higher client engagement and accelerated growth toward its $10 trillion assets-under-management target, though McMillan noted it will take at least a year to measure impact rigorously.

• Meeting AI assistants can improve note quality while freeing advisors to be more present and engaged during client conversations
• Client consent and transparency are essential guardrails when deploying generative AI in high-trust wealth management relationships
• Large-scale enterprise AI rollouts require significant prompt engineering investment — Morgan Stanley spent months fine-tuning Debrief before release