AI in Wealth Management: Beyond the Hype

Ninety-eight percent. That is the share of Morgan Stanley advisor teams using the firm’s AI assistant every single day — across more than 16,000 advisors at a firm not known for moving fast on technology decisions.

Let that number sit for a moment. The industry spent a decade debating whether advisors would adopt AI. The largest wirehouse in the country answered the question while nobody was looking: not as a pilot, not as an experiment — as daily infrastructure. The advisor of 2030 doesn’t compete with AI. She competes with other advisors who have it.

Where AI is actually deployed

Morgan Stanley has gone further than any other major wirehouse in putting AI into the advisor workflow rather than just announcing it. The AI @ Morgan Stanley Assistant, powered by GPT-4 through a partnership with OpenAI dating to 2022, gives advisors conversational access to the firm’s entire intellectual capital library. Debrief came next: consent-based meeting notes, follow-up emails, and action items synced into Salesforce. Jeff McMillan, the firm’s head of firmwide AI, describes the vision as “an efficiency enhancing interaction layer that sits between our colleagues and the many applications they interact with.”

In June 2026, Morgan Stanley went further still — opening its ShareWorks and Equity Edge stock-plan platforms to external AI agents via the Model Context Protocol, an open-source standard for connecting AI systems to data sources. Through its Solium and E-Trade acquisitions, the firm serves stock plans for nearly half of S&P 500 companies; the stated long-term goal is interactions that are “purely agentic.” It is among the earliest examples of a major Wall Street institution letting outside AI agents operate directly on client-facing platforms.

The rest of the industry is not standing still. Merrill and Bank of America Private Bank announced full-scale 2026 rollout of their AI-Powered Meeting Journey — meeting preparation, consent-based summarization, follow-up tasks, documentation. UBS has STAAT in production use, generating pre-meeting briefs and client-opportunity signals. J.P. Morgan Asset Management built Spectrum, a generative AI copilot for portfolio managers that synthesizes proprietary research and filings with retrieval restricted to approved documents; it is in testing with equity managers, with bonds and multi-asset planned.

The most dramatic efficiency numbers come from Asia. Bank of Singapore embedded an agentic AI tool in relationship-manager KYC workflows that extracts and synthesizes records, identifies evidence gaps, and prompts targeted follow-up — cutting average source-of-wealth report preparation from ten days to one hour. Phillip Securities’ POEMSGPT, live since November 2025, answers client questions and interprets financial information inside its POEMS platform, with every response classified as factual or contextual and logged for supervisory review.

Note the pattern: every one of these deployments sits on top of an execution layer. Envestnet’s September 2026 move to add Vestmark’s trading and tax capabilities is the tell — every AI copilot ultimately needs unified managed accounts, tax-loss harvesting, and rebalancing underneath it. AI on the front end is only as good as the pipes behind it.

What AI still can’t do

The honest accounting matters as much as the deployments. AI drafts the follow-up email; it does not sit across from a client in March 2020 and talk them out of selling. It synthesizes research; it does not carry the liability for the recommendation. The planning complexity of a $10 million-plus household — concentrated equity, estate structures, business succession, multigenerational dynamics — remains a human discipline, and the clients who pay premium fees know it.

The demographics argue for augmentation, not replacement. Cerulli projects roughly 26,000 adviser retirements over the decade, with $3.9 trillion in assets expected to change hands. Demand for advice is rising as supply contracts — which is why even aggressive AI forecasts show advisory fee rates holding flat or compressing only modestly, 5 to 10 basis points, in the mass-affluent segment. One scenario analysis has Morgan Stanley’s advisor force declining from 15,000 to 10,000–12,000 over the long term — with substantially higher productivity per advisor. Fewer advisors, each serving larger books at higher satisfaction. That is the bull case, and it is a human-plus-machine story.

The market structure has already rendered its verdict. The pure robo-advisor has not disappeared, but it occupies a defined segment — younger, lower-asset, digitally native investors comfortable with fully automated management — narrower than the original disruption thesis predicted. The winning model is the hybrid: the human advisor using AI as production infrastructure. As Morgan Stanley’s co-head of technology put it, “the companies that are going to survive in the future are the ones who have proprietary data and business logic, which is the foundation of our offering.” Data and judgment are the moat. AI is the lever.

For firms still waiting: the question is no longer whether to adopt, but where the next manual step is. Start there. The copilots are already in production — your competitors’ advisors are using them today.

The category-defining name in wealth management technology — privatewealthtech.com — is currently available for acquisition by the company that will own this space.