Integrating AI into Morgan Stanley Investment Management Research and Investment Workflows
Melvine Manchau
Senior Strategy & Technology Executive | AI & Digital Transformation Leader | Former Salesforce Director | Driving Growth & Innovation in Financial Services | C-Suite Advisor | Product & Program Leadership
August 20, 2025
Morgan Stanley Investment Management (MSIM) has embraced OpenAI-powered tools — notably the AI @ Morgan Stanley Assistant and AI @ Morgan Stanley Debrief — to transform how its teams conduct research, serve clients, and make investment decisions. These generative AI tools are woven into daily workflows across MSIM’s research analysts, portfolio managers, and distribution (sales) teams.
Initially developed within Morgan Stanley’s Wealth Management unit, the Assistant and Debrief are now being scaled firmwide as a “super app” for employees in multiple divisions openai.com. In practice, this means MSIM professionals can use the same AI platforms to quickly retrieve information, synthesize internal data, and even automate routine documentation.
For example, a portfolio manager can query the internal Assistant chatbot to instantly pull insights from Morgan Stanley’s vast research library, while a distribution team member can rely on Debrief to automatically capture notes and action items from a client meeting. Morgan Stanley’s firmwide Head of AI, Jeff McMillan, describes these tools as an “efficiency-enhancing interaction layer” between employees and the many applications they use (from CRM systems to risk analytics) morganstanley.com.
In other words, generative AI is integrated as a connective tissue in MSIM’s processes — bridging the gap between raw information and actionable insight.
Key capabilities enabled by the AI @ Morgan Stanley suite include:
Instant knowledge retrieval:
The Assistant serves as an internal chatbot (powered by GPT-4) that lets MSIM professionals ask complex questions and get answers sourced from the firm’s internal research, market commentary, and intellectual capital emerj.com. It’s been likened to having Morgan Stanley’s smartest expert on call for every employee. Advisors and analysts can pull up data on a company or sector, or glean historical market insights, within seconds — instead of digging through hundreds of thousands of documents manually.
Summarizing research and data:
The generative AI can condense lengthy reports and market data into concise summaries. Morgan Stanley’s Research division, for instance, launched AskResearchGPT in late 2024 to let banking, trading, and MSIM staff query over 70,000 research reports published annually and get synthesized answers morganstanley.com. The GPT-4 model “supercharges” the ability to draw comprehensive insights from multiple documents at once . This capability is invaluable to MSIM’s investment teams, who regularly consume research on economies, industries, and companies. Instead of reading dozens of PDFs, an analyst can ask the AI to “summarize Morgan Stanley’s latest outlook on the semiconductor sector” and receive a cogent digest complete with relevant analyst views — even including hyperlinks to the source reports for deeper follow-up.
Automating meeting notes and follow-ups
The AI @ Morgan Stanley Debrief tool acts as a virtual notetaker and report writer for client meetings. With client consent, Debrief uses OpenAI’s Whisper speech-to-text and GPT-4 to transcribe meeting discussions (for example, an MSIM sales call or portfolio review on Zoom) and generate actionable outputs. After the meeting, Debrief automatically produces a summary of key points, identifies any action items, and even drafts a follow-up email that the advisor or sales representative can edit and send. It also logs the meeting notes into the CRM (such as Salesforce) on the advisor’s behalf . By integrating with MSIM’s client relationship systems, Debrief ensures that no important detail or task falls through the cracks. An internal description of the tool notes that it “takes notes, provides summaries, [and] generates follow-up email drafts” so that advisors can be fully engaged in the conversation without worrying about documentation .
Generating reports and client communications:
Both AI tools can draft polished content, which MSIM teams leverage for efficiency. The Assistant can be prompted to compose a first draft of an investment commentary or a client-ready report on a market trend, drawing from approved internal data. Debrief’s meeting summaries often serve as ready-made client emails or briefing memos. Morgan Stanley even patented a one-click workflow to let employees using AskResearchGPT insert AI findings directly into an email draft with citation, showcasing how seamlessly these tools produce client communication materials. Importantly, all AI-generated content is reviewed by the human advisor/analyst before final use, maintaining necessary oversight
Capabilities in Action: Research Synthesis to Client Service
The real-world impact of these capabilities is evident in MSIM’s day-to-day operations. For research and investment analysis, MSIM portfolio managers use the Assistant as a kind of co-pilot. It can sift through internal research reports, market data, and even prior investment committee notes to answer questions or provide summaries.
According to Morgan Stanley’s AI team, they “went from being able to answer 7,000 questions” with earlier systems to now effectively answering any question across a corpus of 100,000+ documents via the GPT-4 Assistant . This means an investment analyst can quickly obtain, for example, a digest of how various Morgan Stanley analysts across the globe view a certain emerging market — a task that once required reading multiple reports.
The Assistant’s strength is in synthesizing unstructured information from the firm’s knowledge base (research papers, investment strategy notes, market commentary) and presenting it in a concise, conversational form. Morgan Stanley’s content library is enormous, spanning “thousands of papers yearly” on everything from capital markets to industry analysis . By training GPT-4 on this internal trove, MSIM effectively has an AI that distills the firm’s collective intelligence on demand. As Jeff McMillan put it, the AI makes “you as smart as the smartest person in the organization” by giving front-line teams instant access to expert knowledge.
In client service and distribution, the tools streamline how MSIM professionals prepare for and follow up from client engagements. The Debrief tool, especially, has become a cornerstone for client meetings. When an MSIM distribution team member (who might regularly meet institutional clients or financial advisors who distribute MSIM’s funds) sits down for a meeting, Debrief is effectively “in the room” (virtually) taking notes. It captures everything — from a client’s risk concerns to an inquiry about fund performance — and later produces a clean summary and next-step list.
Advisors report that this saves significant time: “It’s saving me about half an hour per meeting just by handling all the note-taking,” one Morgan Stanley advisor noted, which frees up time to focus on decision-making during meetings. Debrief will draft a follow-up email highlighting the portfolio changes or product information discussed, which the MSIM professional can quickly review and send while the discussion is still fresh.
Clients have responded positively to these swift, thorough follow-ups, finding the summaries valuable. In essence, routine tasks that used to take advisors hours — like writing detailed meeting notes or compiling market info for a client memo — now happen within minutes. This enables MSIM’s client-facing teams to provide very timely service. For example, an advisor no longer waits until the end of the day (or week) to type up notes; instead, clients might receive a summary email within hours of a meeting with key points and agreed next steps.
Widespread Adoption by Mid-2025
Morgan Stanley’s investment in these AI tools has led to remarkably rapid and broad adoption, including within MSIM. By mid-2025, usage of the AI Assistant and Debrief is nearly universal among the firm’s advisory and investment teams. In wealth management (where the rollout began), “nearly all advisor teams now use AI tools like the Assistant daily,” achieving over a 98% adoption rate .
This figure, reported in mid-2024, represents essentially full uptake among Morgan Stanley’s ~16,000 financial advisors — a strong proxy for the tool’s value and user-friendliness. Vince Lumia, Head of Morgan Stanley Wealth Management Client Segments, noted that after full deployment, “98% of Financial Advisor teams have adopted the Assistant” as of mid-2024 .
By the following year (2025), as the firm extended these AI capabilities to other business units, uptake remained extremely high. MSIM’s investment professionals were quick to embrace the tools, given the clear productivity benefits. Sal Cucchiara, the CIO for Wealth and Investment Management Technology, confirmed that Morgan Stanley was expanding access to the AI tools beyond just financial advisors to “more of its advisors and other professionals” across the wealth management division. investmentnews.com.
This natural evolution indicates that teams in MSIM — which falls under the Investment Management business — also gained access as the firm moved to leverage generative AI firmwide. Internal memos emphasized that these tools are meant to “boost productivity, expand what teams are capable of, and enhance the value we deliver to clients” across roles. In practice, nearly every MSIM portfolio team and distribution unit had integrated the AI Assistant into their research routines by 2025, and Debrief was commonly used for meeting documentation.
Two years prior, such wide adoption was not a given — advisors tend to be wary of new tech, and Morgan Stanley had piloted the Assistant with a small group. But thanks to intensive training, iterative improvement, and proof of value, usage snowballed. Morgan Stanley’s rigorous evaluation framework for AI ensured the tools were accurate and reliable, which helped earn trust from professionals. The firm carefully tested the AI on real-world use cases (e.g. summarizing a lengthy internal report) and refined outputs with expert feedback before wider rollout.
As a result, once deployed, advisor feedback was overwhelmingly positive and word spread internally. By mid-2025, it’s common to hear MSIM team members refer to the Assistant and Debrief as indispensable parts of their workday. Kaitlin Elliott, Morgan Stanley’s Head of Firmwide Generative AI Solutions, noted that advisors became “more engaged with clients, and follow-ups that used to take days now happen within hours” with these tools in place . This cultural acceptance is reflected in the near-total adoption rates. Generative AI has moved from a novelty to a trusted assistant for Morgan Stanley’s workforce — in the span of just a couple years.
Impact on Productivity, Client Communication, and Decision-Making
The introduction of AI @ Morgan Stanley Assistant and Debrief has meaningfully boosted productivity and enhanced both client communication and investment decision-making within MSIM. Morgan Stanley’s own metrics and executive comments underscore a few key impacts by 2025:
Dramatic time savings and efficiency gains:
Advisors and analysts are saving significant time on information gathering and administrative tasks. Morgan Stanley observed that access to relevant documents jumped from about 20% to 80% — meaning what used to be a needle-in-haystack search through research PDFs is now largely automated.
The Assistant’s ability to retrieve answers from the knowledge base in seconds has been a “game-changer” for efficiency . Vince Lumia highlighted that “AI @ Morgan Stanley Debrief drives immense efficiency in an Advisor’s day-to-day, allowing more time to spend on meaningful engagement with clients”.
By offloading note-taking, email drafting, and document search to AI, highly skilled MSIM personnel can reallocate hours toward higher-value activities (like analyzing investment opportunities or talking to clients). One advisor’s testimonial quantified Debrief as saving 30 minutes per meeting in note-taking alone, which over numerous meetings weekly is a massive productivity boost. Morgan Stanley’s tech leaders frequently describe the AI as freeing up “valuable time” for financial advisors and teams to focus on decision-making and client service.
Faster insights and better-informed decisions:
With AI delivering faster insights, MSIM teams can make decisions with a more comprehensive information backdrop.
The speed at which the Assistant can synthesize research means that before making an investment move or recommending a strategy, MSIM professionals can quickly double-check facts or gather multi-angle analysis that they might have skipped in the past due to time constraints. Jeff McMillan noted that advisors can now engage clients on niche topics spontaneously because “the friction between knowledge and communication has gone to zero”.
In an investment context, that translates to more agile and well-informed decision-making. For example, if a client asks about a specific emerging market trend, an MSIM portfolio specialist can, in real time, consult the Assistant for the latest internal outlook and respond confidently on the spot.
Additionally, by surfacing anomalies and key points from data (as some AI tools do), the technology can nudge human experts toward insights they might otherwise miss. Morgan Stanley’s results align with broader industry observations that AI is helping flag market signals and summarize earnings calls in the asset management process In short, decisions are being made with greater speed and backed by a broader swath of knowledge. advisorhub.com.
Enhanced client communication and satisfaction:
The AI tools have tangibly improved the quality and responsiveness of client communications in MSIM. Clients now often receive follow-up summaries, recommendations, or answers to inquiries much faster than before, which makes the firm’s service stand out.
The content of communications is also more consistent and data-backed, since the AI can include relevant research citations and ensure nothing important is omitted. Advisors report that clients appreciate the thorough meeting recaps and feel more engaged in the process.
By automating routine touchpoints (like summary emails or periodic market updates), advisors have more capacity to proactively reach out to clients with personalized advice — strengthening relationships. Morgan Stanley’s internal surveys have noted extremely high client satisfaction with how advisors address questions, which the firm attributes in part to these AI enhancements. Essentially, AI is helping MSIM deliver “higher quality service in a more effective manner” to clients , as Katy Huberty, Morgan Stanley’s Global Research Director, explained regarding their AI deployments. The human touch remains fundamental, but it’s now augmented by speedy research and polished communications that the AI provides behind the scenes.
Notably, these improvements do not replace human judgment — they elevate it. MSIM professionals still make the final calls on investments and advice. The AI may surface a set of data points or draft an email, but the advisor reviews it, ensuring it aligns with the client’s goals and the firm’s standards.
In this way, the AI acts as a junior analyst or assistant that never tires: crunching data, preparing first drafts, and letting the human experts focus on nuance and strategy. As one asset management chief technologist put it, these AI assistants have become “thought partners” — used even to challenge one’s thinking — rather than simply tools for rote work. Morgan Stanley’s experience by mid-2025 shows that when properly implemented, generative AI can boost productivity and enrich decision-making without sacrificing the human expertise and oversight that are critical in finance.
Complementing Next Best Action and Other AI Systems
Morgan Stanley’s AI @ MS Assistant and Debrief do not exist in isolation — they are designed to complement and enhance other AI and analytics systems in the firm’s toolkit, notably the Next Best Action platform. Next Best Action (NBA) is Morgan Stanley’s long-running AI-driven recommendation engine that analyzes client data to suggest personalized actions or investment ideas for advisors to consider.
Introduced in 2018 and refined with machine learning, NBA became widely adopted (over 90% of Morgan Stanley advisors were using it by 2022). It generates prompts such as: which clients might benefit from a portfolio review, which investment product to discuss given a client’s holdings, or even reminders when a client’s life event might warrant a financial planning conversation. In essence, NBA tells the advisor “what to do next” to better serve a client, based on data patterns and predictive analytics.
What the NBA system traditionally did not do was actually execute those communications or delve into the content of the advice — that remained up to the advisor.
This is where the generative AI tools complement NBA. Morgan Stanley’s Jeff McMillan has indicated that they foresee integrating GPT-4 capabilities with the Next Best Action system. In practice, this integration means an advisor could get a suggested action from NBA (for example, “Client X has excess cash — consider discussing bond fund Y”), and then use the Assistant to instantly pull up a summary of bond fund Y’s latest performance and risks, or even draft an email to Client X explaining why it might be a good fit (drawing on approved research)
The generative AI thus helps the advisor carry out the next best action with high-quality, customized content. Morgan Stanley executives describe AI’s role as a connective layer: “serving as an efficiency enhancing layer that sits between our colleagues and the many applications they interact with — such as … CRMs, reporting tools and risk analysis”. Next Best Action is one of those applications (tied into the CRM), and with the Assistant, an advisor can seamlessly go from an NBA prompt to a well-informed client conversation or message.
Additionally, the Debrief tool feeds back into this loop by logging the outcomes of client meetings (notes and follow-ups) into the CRM.
This enriches the data that NBA analyzes. In other words, if Debrief notes that Client X showed interest in bond fund Y in a meeting, the CRM is updated, and NBA can later factor that into future recommendations (perhaps suggesting a timely follow-up or another related product down the line). The AI tools thus work hand-in-hand: predictive analytics (NBA) suggests “which client, what topic” and generative AI (Assistant/Debrief) helps deliver “the how and the content” of that engagement. Together, they help advisors provide highly personalized service at scale.
Morgan Stanley’s approach is that generative AI enhances the existing AI infrastructure — not replacing it but amplifying its impact. McMillan has called this creating a “flywheel” of future solutions where each AI capability boosts another. For example, the insights from Debrief could eventually inform risk management models (flagging common client concerns) or feed an AI that helps in portfolio execution. By mid-2025, Morgan Stanley is just beginning to unlock these synergies, but the vision is clear: a unified AI ecosystem where a variety of AI tools (predictive, generative, analytical) complement each other to support both employees and clients.
Comparisons to Peer Institutions’ Use of Generative AI
Morgan Stanley’s early move to deploy OpenAI’s GPT models in its workflows set it apart in the financial industry, but it isn’t alone for long. By 2025, many peer institutions in wealth and asset management are also experimenting with or rolling out generative AI solutions, each tailored to their needs. Comparing Morgan Stanley’s approach to others provides context:
Industry-leading adoption:
Morgan Stanley was the first major wealth manager to partner with OpenAI (announced in early 2023) and to deploy a custom GPT-4 assistant firmwide.
This head start is reflected in the 98% adoption rate and daily use by its advisors, which by mid-2025 is a benchmark few others have matched.
Many competitors have been more cautious, running limited pilots. For instance, Bank of America and JPMorgan Chase initially restricted employee use of ChatGPT in 2023, although they later began developing internal AI tools. Morgan Stanley’s willingness to invest early, with robust guardrails (like OpenAI’s data privacy assurances and custom evaluation tests), has given it a lead in real-world AI integration.
Its AI Assistant and Debrief are often cited as exemplars in the industry and have even won innovation awards (e.g. Celent’s Model Wealth Manager Award in 2024 and 2025). Peers are now closely watching Morgan Stanley’s playbook as they craft their own strategies.
Use cases at other asset managers:
Several asset management firms have introduced generative AI into investment processes, though often focusing more on investment research and trading rather than client-facing tools.
For example, J.P. Morgan Asset Management has embedded large language models into its internal platform “Spectrum” to assist portfolio managers in decision-making. One novel function at JPMorgan automatically tracks how long a manager holds a stock versus an optimal timeframe and uses AI to alert if it’s too long or short — essentially nudging a review of the decision.
The AI also helps generate baskets of companies that might gain or lose from certain policies, and flags when an analyst’s forecasts deviate from consensus. This is a more quantitative, investment-strategy use of generative AI, directly aiding portfolio construction and risk monitoring.
Meanwhile, Robeco, a Netherlands-based asset manager, started by using LLMs to draft sections of investor letters but has progressed to more advanced implementations — their AI scans regulatory filings, earnings calls, and even social media to detect emerging investment themes.
Notably, Robeco launched a thematic fund (a UCITS ETF) in late 2024 built around AI-identified themes, showing confidence in AI’s insights driving product designa. AllianceBernstein (AB) is another peer pushing the envelope: AB’s research team has tested “agentic” AI models that autonomously analyze companies and could eventually propose trades within preset boundaries.
The Chief AI Officer at AB uses generative models as a “thought partner” to stress-test his assumptions — illustrating how buy-side firms see AI as a collaborator for human experts. These examples underscore that across the industry, AI is moving from back-office support to front-line decision support, much as Morgan Stanley has done with its advisor-centric tools.
Wealth management and client service at peers:
Other wealth managers have started to follow Morgan Stanley’s lead in client-facing AI. UBS, for instance, has been reportedly exploring GPT-style tools to help its advisors (especially after its integration of Credit Suisse, UBS has a vast global wealth unit eager for efficiency). In the independent advisory segment, firms are also adopting AI: surveys show many registered investment advisors are trying out tools for drafting client emails or parsing product literature.
However, these efforts often rely on third-party fintech solutions or public models with caution. Morgan Stanley’s differentiator is the proprietary, secure nature of its tools — built with OpenAI but trained on its own data and operating behind its firewall. This has alleviated compliance concerns (no client data is exposed publicly, and OpenAI’s models don’t retain MS data) and thus allowed far wider use.
By mid-2025, Morgan Stanley has essentially set a benchmark: a firmwide AI assistant deployed to thousands of front-office personnel. In contrast, many peers are still in the stage of selective rollouts. For example, JPMorgan Chase in 2023 trademarked “IndexGPT,” signaling plans for an AI stock selection or advice engine, but its launch is expected later in 2025. Goldman Sachs has focused AI efforts on automating internal workflows and code generation (Goldman used an in-house LLM called GS Genie for software development tasks), and it is reportedly evaluating generative AI for its research analysts, but it hasn’t given all client advisors a chatbot yet. Wells Fargo and Citigroup have also taken cautious approaches, mainly limiting AI to internal knowledge bases or piloting with small groups of bankers.
Overall, Morgan Stanley’s use of the Assistant and Debrief in MSIM exemplifies a broader trend: generative AI in finance is shifting from hype to real productivity tool. A July 2025 Bloomberg analysis noted that on Wall Street, “large language models now summarize earnings calls, point out market anomalies, draft research, and nudge portfolio decisions” daily, marking “the early stages of a profound shift” in investment workflows.
Morgan Stanley is at the forefront of this shift in the wealth and asset management arena. Its strategy — combining in-house data with advanced AI models and focusing on augmenting human advisors — is being closely watched by competitors.
As more institutions follow suit, we can expect a new industry standard where nearly every asset and wealth manager has a generative AI assistant or co-pilot integrated into their operations. Morgan Stanley Investment Management’s experience by mid-2025 provides a compelling case study of the benefits: faster research synthesis, more engaging client service, greater scale for personalized advice, and significant efficiency gains — all while keeping the human advisor at the center of the client relationship.
Sources: Recent press releases and case studies from Morgan Stanley and OpenAI, industry news articles, and expert commentary have been used to validate these insights.
Notably, Morgan Stanley’s June 2024 announcement of the Debrief tooland OpenAI’s 2024 case study on Morgan Stanley confirm the tools’ capabilities and widespread adoption.
Morgan Stanley’s October 2024 press release on AskResearchGPT details the research synthesis use case. The firm’s leadership quotes — from Vince Lumia to Jeff McMillan — highlight the strategic intent behind these AI deployments.
Industry comparisons are drawn from Bloomberg News reporting (e.g. on JPMorgan, Robeco, AB)and other financial technology analyses. These sources collectively paint a clear picture of how MSIM is leveraging generative AI by mid-2025, and how it stacks up against the broader financial industry’s AI journey.
