When GPT-4 became capable of answering almost any investment question, Morgan Stanley’s Wealth Management division faced the dilemma common to every knowledge-intensive firm: if AI can do a large part of what the expert does, does that make the expert redundant or more powerful? Morgan Stanley’s advantage in wealth management has never been algorithms — it is its roughly 15,000 financial advisors (CNBC, June 2024) who hold decades-long trust relationships with clients, within a firm now serving over 20 million client relationships (Morgan Stanley 2025 Annual Report). The question was whether AI would erode that advantage or multiply it.

Two paths were available. Path A: the robo-advisor route, would use AI to replace advisors with automated recommendations and compete on price and algorithm, treating financial advice as information retrieval. Path B: the route Morgan Stanley chose — would put AI behind every advisor, leaving humans to own the relationship and judgment while AI handled knowledge retrieval and administrative work. The firm stated its position explicitly when it announced the OpenAI partnership in March 2023:

Morgan Stanley built two AI tools on top of OpenAI’s GPT-4, both designed to make advisors better at the human part of their job rather than to replace it.

AI @ Morgan Stanley Assistant (launched September 2023) gives advisors near-instant access to the firm’s “intellectual capital” — a database of roughly 100,000 internal research reports and documents (CNBC, Sept 2023). Built with GPT-4, it lets advisors query the library in natural language and get synthesized answers in seconds (OpenAI). According to the firm, it raised document retrieval efficiency from 20% to 80% (CTO Magazine, citing MS press release).

AI @ Morgan Stanley Debrief (launched June 2024) is a meeting assistant built on GPT-4 and Whisper that, with client consent, takes notes during client meetings, summarizes key points, drafts a follow-up email for the advisor to edit and send, and saves a note into Salesforce (BusinessWire, June 2024). It is built specifically to remove the post-meeting administrative work so advisors can stay focused on the client (OpenAI).

The intended result: advisors spend more time on relationships, strategic advice, and the situations that require human judgment, and less on tasks that AI handles better.

Since those two wealth-management tools, Morgan Stanley has extended the same playbook to other expert groups inside the firm:

AskResearchGPT (launched 2024) brings the Assistant blueprint to the Institutional Securities side — Investment Banking, Sales & Trading and Research. Built on GPT-4, it lets staff search, synthesize and summarize across the more than 70,000 proprietary research reports the firm publishes each year, and export findings straight into draft emails with hyperlinked citations back to the source (Morgan Stanley, Oct 2024).

DevGen.AI (launched January 2026) targets the firm’s engineers. Built on OpenAI’s GPT models, it translates legacy code — languages like Perl — into plain English that developers can use as a basis for rewriting it in modern languages. In its first five months it processed nine million lines of code, saving the firm’s roughly 15,000 developers an estimated 280,000 hours (Entrepreneur, citing WSJ, Jan 2026).

In each case the pattern is the same as with the advisors: the AI absorbs the slow, mechanical retrieval-and-translation work, and the human keeps the judgment.

Morgan Stanley’s competitive moat in wealth management is not data or algorithms, it is human trust that accumulates over years and cannot be replicated by an automated platform. The HI insight is precise: AI is systematically better at information retrieval and administrative work, while humans remain irreplaceable at judgment, relationship, and the emotional dimensions of financial advice, especially for high-net-worth clients making life-defining decisions. Morgan Stanley built AI to own the former so advisors could fully own the latter. This is upskilling rather than reskilling — advisors kept their role and became more capable in it.

The scale comparison is illustrative rather than exact. Globally, robo-advisors manage on the order of $1–2 trillion in assets, with estimates varying widely by source and definition (one 2024 valuation puts the market at about $1.4 trillion) (Market Research Intellect, 2024). Morgan Stanley’s advisors alone oversee roughly $5.8 trillion in advisor-led client assets as of early 2026 (Financial Planning, April 2026), within a firm whose total client assets across Wealth and Investment Management passed $9 trillion at the end of 2025 (Morgan Stanley 2025 Annual Report). The point is not the precise figures but the direction of the bet: the strategy is built to protect and compound the value of the human advisor, not to automate it away.