How To Build Custom AI Solutions

Lessons from Morgan Stanley's Playbook

Resources Included: AI Tool Evaluation & Integration Checklist (Excel) + Sample job description for “AI Enablement Lead”

Morgan Stanley + Open AI

The firms that win with AI won’t be the fastest to adopt, but the fastest to apply it meaningfully.

Generative AI is no longer a novelty. It’s operational, embedded, and reshaping finance at the workflow level.

Morgan Stanley's approach to embedding AI into their wealth management arm offers a blueprint for organisations and teams aiming to build custom AI solutions that actually work.

In this newsletter, we're unpacking their approach and turning it into a practical guide tailored for leaders.

Morgan Stanley partnered with OpenAI to launch an AI assistant (powered by GPT-4) to help their 16,000+ financial advisers access institutional knowledge more efficiently. Key highlights:

• Used proprietary data to train the assistant for tailored relevance
• Embedded into existing systems (CRM, knowledge base) to reduce friction
• Built a rigorous evaluation framework before rollout
• Achieved 98%+ adoption rate among adviser teams

Their success wasn't about adopting AI for AI's sake, it was about creating fit-for-purpose tools aligned with business needs.

How to Build a Custom AI Assistant for Your Team