Who leads AI at Morgan Stanley?

AI at Morgan Stanley is led by Mike Pizzi (Executive Vice President and Global Head of Technology and Operations); Madhu Coimbatore (Global Head of AI Development Platform); Shailesh Gavankar (Head of AI Strategy and Execution). Ownership of AI on the org chart: Firmwide AI team; Co-President Andy Saperstein co-led AI strategy in 2025.

What AI platform does Morgan Stanley use?

Morgan Stanley builds its own AI tools: the AI @ Morgan Stanley Assistant (Sep 2023) and Debrief (Jun 2024) for advisors, AskResearchGPT for the investment bank (2024), DevGen.AI for legacy code (Jan 2025) and client and service agents in wealth management (live by Oct 2026).

Which AI models does Morgan Stanley use?

Morgan Stanley builds on OpenAI models, as OpenAI's only strategic wealth management client from Mar 2023, and Ted Pick said in Apr 2026 that the firm is working with a beta of Anthropic's Claude Mythos.

How many Morgan Stanley employees use AI?

About 98% of Morgan Stanley's roughly 80,000 employees had access to at least one generative AI tool in Jun 2026, and 98% of financial advisor teams had adopted the Assistant by Jun 2024.

How far has AI spread at Morgan Stanley?

Morgan Stanley has reported no business return it credits to AI, for any use case or for the firm as a whole: the AI gains it has measured are counted in time, such as the developer hours DevGen.AI saved by Jun 2025 and the FIXR agents' reconciliation time per book, with no cost or headcount equivalent. Morgan Stanley has officially reported a company-wide rollout, which places it in Piloting: it said in Oct 2024 that close to half of its 80,000 employees used generative AI tools built with OpenAI, while Jeff McMillan owned AI as head of firmwide AI, and in Jun 2026 that about 98% had access to at least one generative AI tool, a count of access rather than use; its 2026 proxy credits Co-President Andy Saperstein with co-leading the AI strategy in 2025. Morgan Stanley falls short of Scaling because it reports no AI-credited return in any function and no firm-level AI figure, and it has not said who succeeded McMillan when he left the role in Feb 2026, so no named owner of a company-wide AI platform reports its results.

Verified Nov 1, 2026

Morgan Stanley

Piloting

Banking
83,728 employees
In-house tools on OpenAI models
~98% of staff with gen AI access
Evident AI Index #9 (2026)

Morgan Stanley deploys AI through tools it builds in-house on OpenAI models, starting with an assistant for financial advisors in Sep 2023 and spreading to research, legacy code, finance operations and, by Oct 2026, live agents that serve wealth clients and service staff. About 98% of its roughly 80,000 employees had access to at least one generative AI tool in Jun 2026. Jeff McMillan's tenure as head of firmwide AI ended in Feb 2026, and Morgan Stanley publishes no AI budget and no return it credits to AI.

morganstanley.com

Their approach to AI

The platform: Morgan Stanley builds its own AI tools: the AI @ Morgan Stanley Assistant (Sep 2023) and Debrief (Jun 2024) for advisors, AskResearchGPT for the investment bank (2024), DevGen.AI for legacy code (Jan 2025) and client and service agents in wealth management (live by Oct 2026).
The models: Morgan Stanley builds on OpenAI models, as OpenAI's only strategic wealth management client from Mar 2023, and Ted Pick said in Apr 2026 that the firm is working with a beta of Anthropic's Claude Mythos.
The reach: About 98% of Morgan Stanley's roughly 80,000 employees had access to at least one generative AI tool in Jun 2026, and 98% of financial advisor teams had adopted the Assistant by Jun 2024.
The catch: Morgan Stanley reports no return it credits to AI: the developer hours DevGen.AI saved carry no cost or headcount figure. Morgan Stanley also publishes no AI budget or weekly usage count, and has not said who succeeded its head of firmwide AI in Feb 2026.

Snapshot

Employees: 83,728; Sep 30, 2026, up 2% in a year; 30,000 work outside the U.S. (2026 letter)
Technology budget: Not disclosed; Information processing and communications expense was $4.42B in 2025, up from $4.09B (Form 10-K)
AI/ML headcount: Not disclosed; 23,000+ technologists and about 16,000 software developers; a Firmwide AI team and a Machine Learning Research team
Who owns AI: Firmwide AI team; Jeff McMillan led it from Mar 2024 to Feb 2026, reporting to the co-presidents; Andy Saperstein co-led AI strategy in 2025 (2026 proxy)
Platform: Built in-house on OpenAI models; Assistant (Sep 2023), Debrief (Jun 2024), AskResearchGPT (2024), DevGen.AI (Jan 2025), client and service agents (Oct 2026)
Models: OpenAI; Anthropic Mythos in beta; OpenAI's only strategic wealth management client from Mar 2023
Reach: ~98% of employees with access; Jun 2026; close to half used OpenAI-based tools in Oct 2024
Use cases: Hundreds; Generative AI use cases (American Banker, Jun 2026); 3,500+ AI tools in wealth management (Feb 2026); DevGen.AI supports 600+ production use cases (Aug 2026)

Key findings

Morgan Stanley put a bespoke GPT-4 assistant in front of its financial advisors in Sep 2023, the first major Wall Street firm to do so according to Jeff McMillan, and 98% of advisor teams had adopted it by Jun 2024. Morgan Stanley said close to half of its 80,000 employees used generative AI tools built with OpenAI in Oct 2024, and that about 98% had access to at least one generative AI tool by Jun 2026.
Jeff McMillan, appointed Morgan Stanley's head of firmwide AI in Mar 2024 and reporting to the co-presidents, held the role until Feb 2026; the firm's Jan 2026 managing director class named Madhu Coimbatore head of the firmwide AI platform and Shailesh Gavankar head of AI strategy and execution.
Morgan Stanley's DevGen.AI, built in-house on OpenAI GPT models and released in Jan 2025, had modernized about 16 million lines of legacy code by Feb 2026, and Morgan Stanley described it in Aug 2026 as a governed generative and agentic AI platform supporting more than 600 production use cases.
Morgan Stanley publishes no AI budget and no return it credits to AI, for any tool or for the firm, and counts the gains it has measured, such as developer hours saved by DevGen.AI, in time rather than cost or headcount; Jed Finn said in Feb 2026 that the wealth management margin could rise a couple hundred basis points if the firm stopped making the investment, and that doing so would cost growth.
Morgan Stanley Wealth Management had a client assist agent and a service assist agent live by Oct 2026, and Ted Pick told analysts on Oct 14, 2026 that the firm, overweight AI in wealth from its early OpenAI work, is now double overweight, eight days after Evident's AI Index moved Morgan Stanley down four places to ninth among banks.

Where Morgan Stanley sits

Stage: Piloting
Morgan Stanley has reported no business return it credits to AI, for any use case or for the firm as a whole: the AI gains it has measured are counted in time, such as the developer hours DevGen.AI saved by Jun 2025 and the FIXR agents' reconciliation time per book, with no cost or headcount equivalent. Morgan Stanley has officially reported a company-wide rollout, which places it in Piloting: it said in Oct 2024 that close to half of its 80,000 employees used generative AI tools built with OpenAI, while Jeff McMillan owned AI as head of firmwide AI, and in Jun 2026 that about 98% had access to at least one generative AI tool, a count of access rather than use; its 2026 proxy credits Co-President Andy Saperstein with co-leading the AI strategy in 2025. Morgan Stanley falls short of Scaling because it reports no AI-credited return in any function and no firm-level AI figure, and it has not said who succeeded McMillan when he left the role in Feb 2026, so no named owner of a company-wide AI platform reports its results.
How the stages are defined
Ownership: Firmwide AI team; Co-President Andy Saperstein co-led AI strategy in 2025
Platform: Hybrid: Tools built in-house on OpenAI models
Models: OpenAI GPT models; Anthropic Claude Mythos in beta
Hosting: Not disclosed for its OpenAI models; Microsoft Azure is a cloud partner since 2021 in a multi-cloud setup, and Morgan Stanley runs its own data centers
Reach: About 98% of employees with access (Jun 2026)

How they deploy

Build the tools, buy the models

Morgan Stanley Wealth Management announced a strategic initiative with OpenAI in Mar 2023, as one of a handful of GPT-4 launch organizations and OpenAI's only strategic client in wealth management with early access to new products. Morgan Stanley builds the tools itself on OpenAI's models: the AI @ Morgan Stanley Assistant answers advisors only from Morgan Stanley content, with links to the source documents, and Jeff McMillan said in Sep 2023 that the bank spent months curating documents and using human experts to test answers.

Wealth management first

Morgan Stanley started generative AI in wealth management, with the Assistant in Sep 2023 and Debrief in Jun 2024, then moved it into Institutional Securities with AskResearchGPT in 2024, into technology with DevGen.AI in Jan 2025 and into finance operations with FIXR agents by Jun 2026. By Oct 2026 Morgan Stanley Wealth Management had a client assist agent and a service assist agent live. Ted Pick said on Oct 14, 2026 that Morgan Stanley was overweight AI in wealth from the early days of its work with OpenAI and is now double overweight.

AI talent in every business

Ted Pick's 2026 shareholder letter calls AI a strategic enterprise priority, says Morgan Stanley is embedding AI talent and resources across each of its businesses, and describes the tools as a productivity enabler that augments human judgment and advice. Morgan Stanley's Firmwide AI team unifies AI efforts across the firm and stewards their use, according to the firm's technology site.

More capacity, not fewer people

Morgan Stanley frames AI as added capacity. Sharon Yeshaya said in Jan 2026 that operations work once checked by two human teams now runs with one human team and one AI team, so teams can do more work on a different cost base. Mark Mitchell, chief product officer of Morgan Stanley at Work, said in Jun 2026 that agentic AI could let the firm expand customer support, plan administration and its wealth funnel without adding thousands of employees, and Ted Pick said in Oct 2026 that wealth management remains a people business.

Agents that clients and advisors approve

Jed Finn said in Feb 2026 that Morgan Stanley is building super agents made of hundreds or thousands of smaller agents to answer client questions, open accounts, move money and eventually build portfolios that advisors approve before use, with client-facing tasks sequenced because of fraud risk. Morgan Stanley Wealth Management's client assist and service assist agents were live by Oct 2026, working within set limits on data access and actions, according to Hina Shamsi, CIO of wealth management technology. Morgan Stanley at Work is also opening its ShareWorks and Equity Edge stock plan platforms to clients' own AI agents through the Model Context Protocol (Jun 2026).

Deployment model: Morgan Stanley runs AI through a Firmwide AI team that unifies the work across the firm, with AI talent embedded in each business (2026 shareholder letter). Wealth Management built the first AI @ Morgan Stanley tools for financial advisors and, through Wealth Management Technology, the client and service agents live by Oct 2026; Morgan Stanley Research, inside Institutional Securities, built AskResearchGPT; and the Technology division under Mike Pizzi built DevGen.AI, runs an AI Development Platform that provides the foundational infrastructure for generative AI across the firm, and runs sandbox environments where teams test new AI models and open-source tools before they reach live systems (Jun 2026). Morgan Stanley's technology site lists Centers of Excellence for AI, data and public cloud alongside an Innovation Office and an Innovation Lab.
Model providers: Morgan Stanley builds on OpenAI. GPT-4 powers the AI @ Morgan Stanley Assistant (Sep 2023) and AskResearchGPT (2024), Debrief is OpenAI-powered (Jun 2024), and DevGen.AI is built on OpenAI's GPT models (American Banker, Jun 2026); CNBC dates the OpenAI relationship to 2022. Ted Pick told analysts on Apr 15, 2026 that Morgan Stanley is working with the beta of Anthropic's Claude Mythos and is permissioned on Claude Mythos Preview, while a Morgan Stanley spokesperson told American Banker the same day that the firm is not part of the group of more than 40 companies Anthropic is sharing Mythos Preview with. Morgan Stanley's Machine Learning Research team publishes on large language models, but Morgan Stanley has not said whether it runs open-weight or self-trained models in production.
Data and infrastructure: Morgan Stanley's advisor Assistant draws on about 100,000 research reports and documents (Sep 2023), and AskResearchGPT searches more than 70,000 research reports a year from inside browsers, Microsoft Teams and Outlook (Oct 2024). Debrief takes meeting notes only with client consent and saves them into Salesforce (Jun 2024). Morgan Stanley's Technology division supports 10 billion transactions and 15 million software builds a year. Morgan Stanley has not disclosed which cloud or data center serves its OpenAI models. Morgan Stanley agreed in Jun 2021 to move workloads onto Microsoft Azure as part of a cloud-first, multi-cloud strategy, its 2025 Form 10-K lists data centers around the world among its properties, and its technology site names public cloud among its Centers of Excellence.

Platforms and internal tools

AI @ Morgan Stanley AssistantSep 2023
The AI @ Morgan Stanley Assistant is a GPT-4 chatbot that gives financial advisors and their support staff quick answers drawn only from Morgan Stanley's intellectual capital, about 100,000 research reports and documents, with links to the sources.
Users
98% of financial advisor teams (Jun 2024)
Models
OpenAI GPT-4
Source
AI @ Morgan Stanley DebriefJun 2024
Debrief takes notes in advisor meetings with client consent, surfaces action items, drafts a follow-up email for the advisor to edit and saves a note into Salesforce.
Models
OpenAI
Source
AskResearchGPTSummer 2024; announced Oct 23, 2024
AskResearchGPT lets Morgan Stanley's investment banking, sales and trading and research staff pull data, insights and summaries from more than 70,000 research reports a year and send findings to clients in an email draft with citations. Staff asked it three times as many questions as the AskResearch tool it augmented, which had run on traditional AI since 2017.
Models
OpenAI GPT-4
Source
DevGen.AIJan 2025
DevGen.AI is an in-house tool that turns decades-old codebases into plain-English specifications and schematic maps that engineers use to rebuild applications in modern languages, and newer models now automate parts of the translation (American Banker, Jun 2026). DevGen.AI came out of a hackathon at the end of 2023 and was granted a patent in Jun 2025 (Business Insider). Morgan Stanley presented DevGen.AI in Aug 2026 as a governed generative and agentic AI platform for code modernization, testing, remediation and operational workflows, supporting more than 600 production use cases.
Users
Thousands of engineers firmwide (Aug 2026)
Models
OpenAI GPT models
Source
AI Development Platform
Morgan Stanley's AI Development Platform, run from the Technology division, provides the foundational infrastructure for generative AI adoption across the firm, according to a 2026 speaker biography of its global head, Madhu Coimbatore.
Source
Client assist and service assist agentsBy Oct 2026
Morgan Stanley Wealth Management's client assist agent, the first version of its client virtual assistant, answers clients' questions in plain language and completes common tasks, and the service assist agent works alongside client service associates, retrieving information and taking actions such as moving money, opening accounts and changing beneficiaries. CIO Dive also described an adviser assist agent for advisers' day-to-day tasks in Oct 2026 without saying whether it is live.
Source
FIXRBy Jun 2026
FIXR is a production system of cooperating agents that analyzes the breaks left after nightly profit and loss calculations across finance, risk, operations and trade capture systems, clears familiar breaks, proposes fixes for others and turns repeated controller decisions into rules.
Source
LeadIQ
LeadIQ uses AI to introduce Morgan Stanley's financial advisors to clients who are interested in advice; Sharon Yeshaya cited it in Jan 2026 as AI on the revenue side, and Morgan Stanley kept investing in it in 2026.
Source
Next Best Action and GenomeBefore Mar 2023
Next Best Action is an internally built AI engine that sends timely, customized messages to clients and prospects under the advisor's guidance, and Genome uses data analytics and machine learning to personalize client communication in the self-directed and workplace channels.
Source
AI model sandboxes
Morgan Stanley runs technology sandbox environments where teams test new AI models and open-source tools before they go into live systems (American Banker, Jun 2026).
Source

Vendors and partners

OpenAIMar 2023
Model provider behind the Assistant, Debrief, AskResearchGPT and DevGen.AI · Morgan Stanley Wealth Management was one of a handful of GPT-4 launch organizations and the only strategic wealth management client with early access to OpenAI's new products.
AnthropicApr 2026
Claude Mythos beta for security · Ted Pick said on Apr 15, 2026 that Morgan Stanley uses a beta of Claude Mythos; a spokesperson said the firm is not in the group of more than 40 companies Anthropic shared Mythos Preview with.
SalesforceJun 2024
CRM where Debrief saves advisor meeting notes
MicrosoftJun 2021
Azure cloud partner; Teams and Outlook host AskResearchGPT for staff · Morgan Stanley agreed in Jun 2021 to move workloads onto Microsoft Azure within a multi-cloud strategy; Morgan Stanley has not said whether its OpenAI models run on Azure.

Leadership

AI, data and platform leaders

Mike Pizzi

Executive Vice President and Global Head of Technology and Operations

Mike Pizzi has led Technology and Operations at Morgan Stanley since Jan 2025, after serving as head of U.S. Banks and head of Technology from Jan 2023 and as CEO of E*TRADE Financial when Morgan Stanley bought it in 2020. Pizzi oversees about 16,000 software developers, American Banker ranked him 27th among its Most Innovative People in Finance in Jun 2026 for the DevGen.AI work, and he said in Oct 2025 that AI coding has had a profound impact on software development.

Madhu Coimbatore

Global Head of AI Development Platform

Unconfirmed

Madhu Coimbatore was named a managing director in Morgan Stanley's Jan 2026 class as head of the firmwide AI platform, according to eFinancialCareers. A 2026 speaker biography describes Coimbatore as a managing director in the Technology division and global head of the AI Development Platform, responsible for the infrastructure behind generative AI adoption across the firm, who joined Morgan Stanley's technology analyst program in 2007. The current role has not been confirmed by Morgan Stanley.

Shailesh Gavankar

Head of AI Strategy and Execution

Unconfirmed

Shailesh Gavankar was named a managing director in Morgan Stanley's Jan 2026 class as head of AI strategy and execution, according to eFinancialCareers; Morgan Stanley has not published the role itself.

Hina Shamsi

CIO of Wealth Management Technology

Hina Shamsi is the CIO of wealth management technology at Morgan Stanley, whose team runs the client assist and service assist agents. Shamsi said in Oct 2026 that the agents work within defined boundaries on data, actions and human hand-offs, and stressed guardrails, permissions, monitoring and human oversight.

Katy Huberty

Global Head of Research and Senior Relationship Management

Katy Huberty was Morgan Stanley's global director of research and co-chair of its AI Steering Committee when AskResearchGPT launched in Oct 2024, and spoke for the tool; Morgan Stanley described her as global head of research and senior relationship management in Jul 2026. Huberty's current role on the AI Steering Committee has not been confirmed since 2024.

Trevor Brosnan

Global Head of Technology Strategy, Architecture and Modernization

Unconfirmed

Trevor Brosnan led the development of DevGen.AI, according to Business Insider, and said with a colleague in mid-2025 that the tool would not take software engineering jobs; his title has not been confirmed since.

Executives

Ted Pick

Chairman and Chief Executive Officer

Ted Pick became CEO of Morgan Stanley in Jan 2024 and chairman in Jan 2025, after serving as Co-President and co-head of corporate strategy from 2021 to 2023 and head of Institutional Securities from 2018 to 2023. Pick's 2026 shareholder letter calls AI a strategic enterprise priority, and he told analysts in Oct 2026 that AI's evolution is absolutely the top priority for the firm.

Andy Saperstein

Co-President; responsible for Wealth and Investment Management

Andy Saperstein has been Co-President of Morgan Stanley since Jun 2021 and ran Wealth Management from 2019 to 2023. Morgan Stanley's 2026 proxy credits Saperstein with co-leading the firm's AI strategy in 2025, including generative AI capabilities in Wealth Management and Investment Management, and he was quoted announcing the OpenAI initiative in Mar 2023. Jeff McMillan reported to Saperstein and Dan Simkowitz when he became head of firmwide AI in Mar 2024.

Dan Simkowitz

Co-President; responsible for Institutional Securities

Dan Simkowitz became Co-President of Morgan Stanley in Jan 2024 after running Investment Management from 2015 to 2023. Morgan Stanley's 2026 proxy lists AI enhancements as part of Simkowitz's integrated investment bank strategy in 2025.

Sharon Yeshaya

Executive Vice President and Chief Financial Officer

Sharon Yeshaya has been CFO of Morgan Stanley since Jun 2021 and was head of investor relations from 2016 to 2021. Morgan Stanley's 2026 proxy credits Yeshaya with modernizing key finance processes, including through AI, and she describes the firm's AI work on earnings calls, from LeadIQ on the revenue side to AI checks in operations and reconciliations in finance.

Jed Finn

Head of Wealth Management

Jed Finn has headed Morgan Stanley Wealth Management, including E*TRADE and Morgan Stanley at Work, since Jan 2024, after serving as its chief operating officer. Finn described the firm's plan for AI super agents at the UBS Financial Services Conference in Feb 2026, counting more than 3,500 AI-driven tools and capabilities in the business.

Eric Grossman

Chief Legal Officer and Chief Administrative Officer

Eric Grossman has been Morgan Stanley's Chief Legal Officer since Jan 2012 and Chief Administrative Officer since Jul 2022, with responsibility for Legal, Compliance, Global Financial Crimes and Operational Risk. Morgan Stanley's 2026 proxy credits Grossman with promoting AI use across all the areas he covers in 2025 to gain efficiencies.

Board

Jami Miscik

Independent director; Chair of the Operations and Technology Committee

Jami Miscik has been a Morgan Stanley director since 2014 and chairs the board's Operations and Technology Committee, which oversees the firm's technology strategy, including AI, and reviews the technology budget. Miscik is CEO of Global Strategic Insights, a former CEO and vice chair of Kissinger Associates and a former Deputy Director for Intelligence at the CIA.

AI org map

Firmwide AI team with AI talent embedded in each business; tools built by the businesses and Technology

Morgan Stanley named Jeff McMillan its first head of firmwide AI in Mar 2024, reporting to Co-Presidents Andy Saperstein and Dan Simkowitz, with Katy Huberty as co-chair of the firm's AI Steering Committee; the 2026 proxy credits Saperstein with co-leading the AI strategy in 2025. McMillan held the role until Feb 2026, and Morgan Stanley's Jan 2026 managing director class named Madhu Coimbatore head of the firmwide AI platform and Shailesh Gavankar head of AI strategy and execution. Mike Pizzi's Technology and Operations division builds firmwide engineering tools such as DevGen.AI and the AI Development Platform, while Wealth Management and Research build their own AI tools.

Morgan Stanley does not publish an AI or machine learning headcount. Morgan Stanley's Technology division counts more than 23,000 technologists worldwide (technology site, Nov 2026), and Mike Pizzi oversees about 16,000 software developers (American Banker, Jun 2026). A Firmwide AI team unifies AI work across the firm, an AI Development Platform team in Technology builds the infrastructure for generative AI, a Machine Learning Research team of specialist researchers works on problems in fixed income, investment management and electronic trading, and Ted Pick's 2026 letter says AI talent and resources are embedded in each business.

Reporting lines

Chairman and CEO
Ted Pick
- Co-President, Wealth and Investment Management
  Andy Saperstein
  Co-led the firm's AI strategy in 2025 (2026 proxy)
  - Wealth Management
    Jed Finn
    AI @ Morgan Stanley tools, LeadIQ, client and service agents and super agents; the line to Andy Saperstein is inferred from his responsibilities
    - Morgan Stanley at Work product
      Mark Mitchell, Chief Product Officer
      Opening stock plan platforms to client AI agents through the Model Context Protocol
- Co-President, Institutional Securities
  Dan Simkowitz
  - Morgan Stanley Research
    Katy Huberty, Global Head of Research
    Built AskResearchGPT; the reporting line is inferred, not published
- Chief Financial Officer
  Sharon Yeshaya
- Legal and Administration
  Eric Grossman
  Legal, Compliance and Operational Risk; promoted AI use in 2025 (2026 proxy)
- Technology and Operations
  Mike Pizzi
  23,000+ technologists; DevGen.AI, AI sandboxes and Centers of Excellence for AI, data and public cloud
  - AI Development Platform
    Madhu Coimbatore
    Infrastructure for generative AI across the firm; listed as the firmwide AI platform in the Jan 2026 managing director list
  - Wealth Management Technology
    Hina Shamsi, CIO
    Client assist and service assist agents; placement inferred from the title
  - Technology Strategy, Architecture and Modernization
    Trevor Brosnan
    Led DevGen.AI (Jul 2025); title unconfirmed since
Firmwide AI team
Head of firmwide AI Jeff McMillan reported to the two co-presidents from Mar 2024 and left the role in Feb 2026; Morgan Stanley has not published where the team now reports
- AI Strategy and Execution
  Shailesh Gavankar
  Placement inferred from the title

Morgan Stanley's 2025 Form 10-K lists Ted Pick, Eric Grossman, Mike Pizzi, Andy Saperstein, Dan Simkowitz and Sharon Yeshaya among its executive officers, but no AI executive. Morgan Stanley has not published who now holds the head of firmwide AI title after Jeff McMillan, or to whom Madhu Coimbatore and Shailesh Gavankar report.

Where the AI workforce sits

Technology division
Morgan Stanley's technology site, Nov 2026; about 16,000 of them are software developers (American Banker, Jun 2026). Not all work on AI.
23,000+ technologists
Firmwide AI team and Machine Learning Research
Morgan Stanley has not published the size of its central AI and research teams.
Not disclosed
AI talent in the businesses
Ted Pick's 2026 letter says AI talent and resources are embedded across each business; no count is given.
Not disclosed

Named people

Executive sponsors

Ted Pick
Chairman and CEO
Andy Saperstein
Co-President; responsible for Wealth and Investment Management · Co-led the firm's AI strategy in 2025 (2026 proxy)
Dan Simkowitz
Co-President; responsible for Institutional Securities · AI enhancements in the investment bank (2026 proxy)
Sharon Yeshaya
Chief Financial Officer · Modernizing finance processes, including through AI (2026 proxy)
Eric Grossman
Chief Legal Officer and Chief Administrative Officer · Promoted AI use across Legal and the risk functions he covers (2026 proxy)

Firmwide AI

Jeff McMillan was head of firmwide AI from Mar 2024 until Feb 2026; both titles below come from the Jan 2026 managing director list and are unconfirmed by Morgan Stanley.

Madhu Coimbatore
Global Head of AI Development Platform
Shailesh Gavankar
Head of AI Strategy and Execution

Technology and Operations

23,000+ technologists; about 16,000 software developers (2026)

Mike Pizzi
Global Head of Technology and Operations
Trevor Brosnan
Global Head of Technology Strategy, Architecture and Modernization · Led DevGen.AI (Jul 2025); title unconfirmed since
Todd Johnson
Managing Director · Presented the FIXR reconciliation agents (Jun 2026)
Hina Shamsi
CIO of Wealth Management Technology · Client assist and service assist agents (Oct 2026)

Business-line AI

Jed Finn
Head of Wealth Management · Super agents and AI @ Morgan Stanley tools for advisors
Katy Huberty
Global Head of Research and Senior Relationship Management · AskResearchGPT; co-chair of the AI Steering Committee in Oct 2024
Mark Mitchell
Chief Product Officer, Morgan Stanley at Work · Opening ShareWorks and Equity Edge to client AI agents (Jun 2026)

Use cases

Morgan Stanley runs generative AI across hundreds of use cases, American Banker reported in Jun 2026, Jed Finn counted more than 3,500 AI-driven tools and capabilities in wealth management in Feb 2026, and client assist and service assist agents were live in wealth management by Oct 2026. Morgan Stanley has published no return it credits to AI for any of these tools: DevGen.AI's saving is counted in developer hours and the FIXR reconciliation agents' in time per book, neither with a cost or headcount figure. Morgan Stanley gives no firm-level return for AI; Sharon Yeshaya said in Jan 2026 that some of its AI work, such as LeadIQ, sits on the revenue side, without giving a figure.

Use case	Area	Status
AI @ Morgan Stanley Assistant Financial advisors and their teams use the GPT-4 Assistant for quick access to Morgan Stanley's intellectual capital, with answers generated only from Morgan Stanley content; 98% of advisor teams had adopted it by Jun 2024.	Wealth Management	In production
AI @ Morgan Stanley Debrief Debrief writes notes in client meetings with consent, surfaces action items, drafts the follow-up email and saves a note to Salesforce.	Wealth Management	In production
LeadIQ advisor matching LeadIQ uses AI to introduce advisors to clients who are interested in advice, which Sharon Yeshaya described as AI on the revenue side.	Wealth Management	In production
Client assist agent The client assist agent, the first version of Morgan Stanley's client virtual assistant, lets wealth clients ask questions in plain language and complete common tasks; Morgan Stanley plans to make it more account-specific.	Wealth Management	In production
Service assist agent The service assist agent works alongside Morgan Stanley's client service associates, retrieving information and taking actions such as moving money, opening accounts and changing beneficiaries within set permissions.	Wealth Management	In production
Super agents for advisors and clients Morgan Stanley is building super agents from many smaller agents to answer client questions, open accounts, move money, draft tax plans and build portfolios that advisors approve, starting with balances and tax documents for clients.	Wealth Management	Announced
Client AI agents on Morgan Stanley at Work A handful of corporate clients have early access for their AI agents to pull data from ShareWorks and Equity Edge through the Model Context Protocol, and Morgan Stanley plans to open access to its 3,400 administration clients by 2027.	Wealth Management	Pilot
AskResearchGPT AskResearchGPT answers investment banking, sales and trading and research staff from more than 70,000 research reports a year; Katy Huberty said a salesperson answers the average client inquiry in one-tenth of the time.	Institutional Securities	In production
DevGen.AI legacy code modernization DevGen.AI turns old code into specifications and schematic maps for rebuilding systems behind trading, wealth management and client platforms, and had modernized about 16 million lines by Feb 2026. Trevor Brosnan said DevGen.AI had saved developers more than 280,000 hours by Jun 2025, and Morgan Stanley said the time saved would not cost engineering jobs (Business Insider, Jul 2025).	Technology	In production
AI coding for developers Mike Pizzi said in Oct 2025 that AI has already changed how Morgan Stanley's coders work, producing more code and leaving engineers more time for work such as code reviews.	Technology	In production
FIXR profit and loss reconciliation agents FIXR agents work through the nightly breaks between finance, risk, operations and trade capture systems, cutting the time to reconcile a book from up to six hours to two or three, with a controller reviewing every recommendation.	Finance and operations	In production
AI checks on operations documentation Sharon Yeshaya said in Jan 2026 that Morgan Stanley operations work once done by two human teams checking each other's documentation now runs with one human team and one AI team.	Finance and operations	In production
Vulnerability detection with Claude Mythos Morgan Stanley is working with a beta version of Anthropic's Claude Mythos, a model that detects security vulnerabilities in software, and is looking at places in its infrastructure to use it, Ted Pick said in Apr 2026.	Cybersecurity	Pilot

Adoption

Morgan Stanley spread AI from wealth management outward. Machine learning tools such as Next Best Action and the AskResearch chatbot, in use since 2017, came first, then the GPT-4 Assistant for financial advisors in Sep 2023, adopted by 98% of advisor teams by Jun 2024. Close to half of Morgan Stanley's 80,000 employees used generative AI tools built with OpenAI in Oct 2024, by Morgan Stanley's own count, while Jeff McMillan led AI as head of firmwide AI, and about 98% had access to at least one generative AI tool by Jun 2026. Morgan Stanley has not published a weekly or daily active user count, so how many of those with access use the tools regularly is not known.

By the numbers

Employees with generative AI access: ~98%; Of about 80,000 employees, per Morgan Stanley to American Banker, Jun 2026

AI tools and capabilities in wealth management: 3,500+; Jed Finn, head of wealth management, Feb 2026

Legacy code modernized by DevGen.AI: ~16 million lines; As of Feb 2026 (American Banker, Jun 2026)

Evident AI Index for banks: #9; Oct 2026 edition, down four places

Technologists: 23,000+; Morgan Stanley technology page, Nov 2026; about 16,000 software developers (Jun 2026)

Employees: 83,728; Sep 30, 2026, Q3 2026 earnings release

Rollout, first pilot to today

2017
Morgan Stanley Research began using AskResearch, a chatbot based on traditional AI that later became the base for AskResearchGPT.
CNBC
Mar 14, 2023
Morgan Stanley Wealth Management announced a strategic initiative with OpenAI to build a GPT-4 assistant on its own content, alongside its Next Best Action engine and Genome machine learning tools.
Morgan Stanley
Sep 18, 2023
The AI @ Morgan Stanley Assistant went fully live for financial advisors and their support staff, with access to about 100,000 research reports and documents.
CNBC
Mar 2024
Morgan Stanley named Jeff McMillan, previously chief analytics and data officer, its head of firmwide AI.
CDO Magazine
Jun 26, 2024
Morgan Stanley launched AI @ Morgan Stanley Debrief and said 98% of financial advisor teams had adopted the Assistant.
Morgan Stanley
Oct 23, 2024
Morgan Stanley announced AskResearchGPT, rolled out to Institutional Securities from summer 2024, and said close to half of its 80,000 employees were using generative AI tools built with OpenAI.
CNBC
Jan 2025
Morgan Stanley released DevGen.AI internally to translate legacy code into specifications for modern languages.
Business Insider via AOL
Feb 2026
Jeff McMillan left Morgan Stanley's head of firmwide AI role, which he had held since Mar 2024.
American Banker
Jun 2026
Morgan Stanley said about 98% of its roughly 80,000 employees had access to at least one generative AI tool, across hundreds of use cases.
American Banker
Aug 18, 2026
Morgan Stanley presented DevGen.AI at Foundry's CIO 100 event as a governed generative and agentic AI platform supporting more than 600 production use cases, used firmwide by thousands of engineers.
Foundry
Oct 2026
Morgan Stanley Wealth Management had its client assist agent and service assist agent live, the first agents serving its wealth clients and client service associates directly.
CIO Dive

Economics & controls

Spend and ROI

Technology spend

Morgan Stanley does not publish a technology budget. Morgan Stanley's information processing and communications expense rose to $4.42 billion in 2025 from $4.09 billion in 2024 (2025 Form 10-K), and the firm said non-compensation expenses in the third quarter of 2026 rose partly because of higher technology spend. Sharon Yeshaya said on Jul 15, 2026 that higher technology-driven spend relates to investments in infrastructure and AI-enabled efficiencies.

What AI itself costs

Morgan Stanley has not disclosed what it spends on AI. Jed Finn said in Feb 2026, while describing the firm's AI super agents, that Morgan Stanley could lift its wealth management margin a couple hundred basis points if it stopped making the investment, but that doing so would sacrifice growth. Asked about AI spending on Oct 14, 2026, Sharon Yeshaya said companies are using their existing technology dollars and expense base to work out how AI can make them more productive, without giving a Morgan Stanley figure, and Ted Pick said Morgan Stanley, overweight AI in wealth from its early OpenAI work, is now double overweight.

Reported ROI

Morgan Stanley has not published a return it credits to AI, for the firm or for any tool. The AI gains Morgan Stanley has put numbers on are counted in time: Trevor Brosnan said DevGen.AI had saved developers more than 280,000 hours by Jun 2025, with no cost or headcount equivalent given. VentureBeat put the saving from the FIXR reconciliation agents at about 1,500 controller hours a week (Jun 2026), but attributed only the per-book timings to Morgan Stanley's Todd Johnson, not the weekly total. Sharon Yeshaya said in Jan 2026 that some of the firm's AI work, such as LeadIQ, sits on the revenue side, without giving a figure, and Ted Pick said in Apr 2026 that the efficiency and effectiveness gains from AI should not be dismissed.

Productivity statements

Ted Pick said in Jan 2026 that Morgan Stanley can get more operating efficiency out of its infrastructure through AI, and Sharon Yeshaya said the firm, as a first adopter, is already seeing productivity gains, such as operations teams where one AI team now checks documentation alongside one human team. Speaking as head of finance in Oct 2026, Yeshaya said AI is cutting manual hours in reconciliations across books and helping turn spreadsheet-based processes into code that moves into production, and that equities salespeople have AI tools. Katy Huberty said in Oct 2024 that salespeople answer the average client inquiry in one-tenth of the time with AskResearchGPT. Mike Pizzi said in Jun 2026 that AI will lead Morgan Stanley to develop more software, so projects that did not clear the bar before can go ahead.

Headcount

Morgan Stanley had 83,728 employees on Sep 30, 2026, up 2% from 82,398 a year earlier, and Ted Pick's 2026 letter counts 30,000 of its 83,000 employees outside the United States. Mark Mitchell, chief product officer of Morgan Stanley at Work, said in Jun 2026 that agentic AI could let the firm expand customer support, plan administration and its wealth funnel without adding thousands of employees. Morgan Stanley has not tied a headcount reduction to AI in its 2025 Form 10-K or its 2026 earnings calls: it cut about 2,500 jobs, about 3% of staff, in Mar 2026 in what Bloomberg called a broader realignment, the first quarter of 2026 carried $178 million of severance charges with no link to AI given, and the DevGen.AI team said in mid-2025 that the tool would not take engineering jobs.

Governance and risk

Board oversight

Morgan Stanley's Operations and Technology Committee of the board, chaired by Jami Miscik, oversees the firm's operations and technology strategy, including trends such as AI, and reviews the technology budget and major technology investments (2026 proxy). Morgan Stanley's board also schedules in-depth sessions on technology initiatives, including AI, according to the same proxy.

Model risk review

Morgan Stanley's Model Risk Management Department, inside the independent Risk Management function, validates and approves models for use, recertifies them every year, tracks fixes for model limitations and oversees a firmwide model inventory, according to the 2025 Form 10-K. Morgan Stanley has not said how generative AI models and agents fall under that review.

Answers grounded in firm content

Morgan Stanley built the AI @ Morgan Stanley Assistant to answer only from Morgan Stanley Wealth Management content, with links to source documents, and tested its answers with human experts for months before launch (Sep 2023). Morgan Stanley's Debrief takes notes only with client consent, and AskResearchGPT answers carry citations back to the original research.

Human accountability for agents

Morgan Stanley's FIXR agents ask for help when unsure, and a controller reviews, approves or corrects every recommendation; Todd Johnson said in Jun 2026 that a strong principle of the firm's AI governance is that there always has to be human accountability. Jed Finn said in Feb 2026 that advisors will approve portfolio recommendations from super agents before they are carried out, and that client-facing agents will move money only after simpler tasks, because of fraud risk. Hina Shamsi, CIO of wealth management technology, said in Oct 2026 that the wealth agents work within clearly defined limits on what data they reach, what actions they take and when a human must step in, and that agents should not have unlimited autonomy.

Testing before production

Morgan Stanley runs technology sandboxes in which teams test new AI models and open-source tools before they go into live systems (American Banker, Jun 2026).

Cyber defense

Morgan Stanley's 2025 Form 10-K warns that attackers may use AI to make fraud and intrusion more effective and that reliance on generative AI can heighten information security risk. Ted Pick said on Apr 15, 2026 that Morgan Stanley is working with a beta of Anthropic's Claude Mythos and would take its defenses to another level if the model raised cyber risk across the system, the week after he and other bank chief executives met Anthropic executives, the Federal Reserve chair and the Treasury secretary in Washington.

Disclosed risks

Morgan Stanley's 2025 Form 10-K lists generative AI among the technological risks to its results, says generative AI and tokenization will likely keep pressure on Institutional Securities revenues, and names laws on AI among the regulations that could affect profitability.

Recent signals

Oct 26, 2026
CIO Dive reported that Morgan Stanley Wealth Management has a client assist agent and a service assist agent live, with Hina Shamsi, CIO of wealth management technology, stressing guardrails, permissions, monitoring and human oversight.
CIO Dive
Oct 14, 2026
Ted Pick said on the Q3 2026 call that Morgan Stanley, overweight AI in wealth from its early OpenAI work, is now double overweight, that AI's evolution is absolutely the top priority for the firm, and that wealth management remains a people business; Sharon Yeshaya said the firm keeps embracing AI tools across its technology stack.
Q3 2026 earnings call transcript
Oct 6, 2026
Evident's 2026 AI Index for banks ranked Morgan Stanley ninth, down four places from the year before.
Evident
Aug 18, 2026
Morgan Stanley showcased DevGen.AI at Foundry's CIO 100 event as a governed generative and agentic AI platform with more than 600 production use cases, used firmwide by thousands of engineers.
Foundry
Jul 15, 2026
Sharon Yeshaya said on the Q2 2026 call that higher technology-driven spend at Morgan Stanley funds infrastructure and AI-enabled efficiencies, and that investment is going into LeadIQ to match individuals with advisors.
Q2 2026 earnings call transcript
Jun 30, 2026
VentureBeat reported that Morgan Stanley's FIXR agents cut the time to reconcile a trading book's profit and loss from up to six hours to two or three, with a human approving every recommendation.
VentureBeat
Jun 3, 2026
CNBC reported that Morgan Stanley at Work will let clients' AI agents connect to ShareWorks and Equity Edge through the Model Context Protocol, with access for its 3,400 administration clients planned by 2027.
CNBC
Jun 2, 2026
Brillio named Jeff McMillan, former head of firmwide AI at Morgan Stanley and founder of the advisory firm McMillanAI, an independent director.
Brillio via PR Newswire
Jun 1, 2026
American Banker reported that about 98% of Morgan Stanley's 80,000 employees have access to at least one generative AI tool and that DevGen.AI had modernized about 16 million lines of code by Feb 2026.
American Banker
Apr 15, 2026
Ted Pick said on the Q1 2026 call that Morgan Stanley is working with the beta of Anthropic's Claude Mythos and that AI will grow from automating routine tasks into a productivity phenomenon; Sharon Yeshaya said the firm is investing in its agentic infrastructure.
Q1 2026 earnings call transcript
Apr 2, 2026
Morgan Stanley filed its 2025 Annual Report, in which Ted Pick's shareholder letter calls AI a strategic enterprise priority and says the firm is embedding AI talent and resources across each of its businesses, alongside a 2026 proxy crediting Andy Saperstein with co-leading the AI strategy in 2025.
2025 Annual Report
Feb 10, 2026
Jed Finn said at the UBS Financial Services Conference that Morgan Stanley is building AI super agents for advisors and clients and has more than 3,500 AI-driven tools and capabilities in wealth management.
Financial Planning
Jan 15, 2026
Sharon Yeshaya said on the Q4 2025 call that LeadIQ puts some of Morgan Stanley's AI on the revenue side and that some operations teams now pair one human team with one AI team.
Q4 2025 earnings call transcript
Jan 11, 2026
Morgan Stanley's 2026 managing director class included Madhu Coimbatore, head of the firmwide AI platform, and Shailesh Gavankar, head of AI strategy and execution, eFinancialCareers reported.
eFinancialCareers

Quick answers

How does Morgan Stanley deploy AI?: Morgan Stanley deploys AI through tools it builds in-house on OpenAI models, starting with an assistant for financial advisors in Sep 2023 and spreading to research, legacy code, finance operations and, by Oct 2026, live agents that serve wealth clients and service staff. About 98% of its roughly 80,000 employees had access to at least one generative AI tool in Jun 2026. Jeff McMillan's tenure as head of firmwide AI ended in Feb 2026, and Morgan Stanley publishes no AI budget and no return it credits to AI.
What does Morgan Stanley use AI for?: Morgan Stanley runs generative AI across hundreds of use cases, American Banker reported in Jun 2026, Jed Finn counted more than 3,500 AI-driven tools and capabilities in wealth management in Feb 2026, and client assist and service assist agents were live in wealth management by Oct 2026. Morgan Stanley has published no return it credits to AI for any of these tools: DevGen.AI's saving is counted in developer hours and the FIXR reconciliation agents' in time per book, neither with a cost or headcount figure. Morgan Stanley gives no firm-level return for AI; Sharon Yeshaya said in Jan 2026 that some of its AI work, such as LeadIQ, sits on the revenue side, without giving a figure.
Who leads AI at Morgan Stanley?: AI at Morgan Stanley is led by Mike Pizzi (Executive Vice President and Global Head of Technology and Operations); Madhu Coimbatore (Global Head of AI Development Platform); Shailesh Gavankar (Head of AI Strategy and Execution). Ownership of AI on the org chart: Firmwide AI team; Co-President Andy Saperstein co-led AI strategy in 2025.
What AI platform does Morgan Stanley use?: Morgan Stanley builds its own AI tools: the AI @ Morgan Stanley Assistant (Sep 2023) and Debrief (Jun 2024) for advisors, AskResearchGPT for the investment bank (2024), DevGen.AI for legacy code (Jan 2025) and client and service agents in wealth management (live by Oct 2026).
Which AI models does Morgan Stanley use?: Morgan Stanley builds on OpenAI models, as OpenAI's only strategic wealth management client from Mar 2023, and Ted Pick said in Apr 2026 that the firm is working with a beta of Anthropic's Claude Mythos.
How many Morgan Stanley employees use AI?: About 98% of Morgan Stanley's roughly 80,000 employees had access to at least one generative AI tool in Jun 2026, and 98% of financial advisor teams had adopted the Assistant by Jun 2024.
How far has AI spread at Morgan Stanley?: Morgan Stanley has reported no business return it credits to AI, for any use case or for the firm as a whole: the AI gains it has measured are counted in time, such as the developer hours DevGen.AI saved by Jun 2025 and the FIXR agents' reconciliation time per book, with no cost or headcount equivalent. Morgan Stanley has officially reported a company-wide rollout, which places it in Piloting: it said in Oct 2024 that close to half of its 80,000 employees used generative AI tools built with OpenAI, while Jeff McMillan owned AI as head of firmwide AI, and in Jun 2026 that about 98% had access to at least one generative AI tool, a count of access rather than use; its 2026 proxy credits Co-President Andy Saperstein with co-leading the AI strategy in 2025. Morgan Stanley falls short of Scaling because it reports no AI-credited return in any function and no firm-level AI figure, and it has not said who succeeded McMillan when he left the role in Feb 2026, so no named owner of a company-wide AI platform reports its results.

Open questions

Who holds Morgan Stanley's head of firmwide AI title after Jeff McMillan left it in Feb 2026, do Madhu Coimbatore and Shailesh Gavankar still hold the roles named in Jan 2026, and to whom do they report?
Which cloud or data center serves the OpenAI models behind Morgan Stanley's tools, and on what data retention terms?
How much does Morgan Stanley spend on AI, and what is its total technology budget?
How many of the roughly 98% of Morgan Stanley employees with access to a generative AI tool use one every week, against the close to half who used one in Oct 2024?
What revenue or cost return do the AI @ Morgan Stanley Assistant, Debrief, AskResearchGPT and LeadIQ deliver, and what cost or headcount saving does Morgan Stanley itself count from the FIXR agents?
What productivity gains has Morgan Stanley measured from its client assist and service assist agents, and how many clients use them?
How many of the 3,500 AI-driven tools and capabilities in Morgan Stanley's wealth business are in production, and what counts as one?
What cost or headcount equivalent does Morgan Stanley put on the more than 280,000 developer hours DevGen.AI had saved by Jun 2025, and how many has DevGen.AI saved since?
Do generative AI models and agents go through Morgan Stanley's Model Risk Management validation, and on what terms?
What did Morgan Stanley find in its beta of Claude Mythos, and does it use other Anthropic models?
When will Morgan Stanley's super agents launch, and will all 3,400 Morgan Stanley at Work administration clients get agent access in 2027?
Why did Morgan Stanley fall four places to ninth in Evident's 2026 AI Index?

Sources

Morgan Stanley

Filings, letters and investor materials

Indexes, events and partners

News

Data confidence

Confirmed: Stated in a primary or reputable public source.
Estimated: Derived or approximated from public reporting.
Unconfirmed: Reported, but not yet corroborated.

Morgan Stanley profile · Enterprise AI diffusion report by Imagine · facts checked against their sources on Nov 1, 2026

More company profiles

All profiles