The most revealing question a wealth management CEO can ask about artificial intelligence is what advisers are doing differently on Monday morning.
Are they having better conversations? Identifying needs they previously missed? Following through more consistently? Serving more clients without weakening the quality of advice?
Those questions deserve as much attention as the technology itself.
From my experience in practice management, a substantial gap can exist between introducing a capability and changing how people work. AI makes that gap particularly consequential. Institutions can invest heavily, announce impressive deployments and still struggle to demonstrate what has improved for clients or the business.
I believe this creates an important opportunity for experienced managers. People who understand advisers, client relationships and practice economics have a central role in making AI work. To fulfil it, they also need a practical understanding of what the technology can do.
Much of the visible progress so far concerns productivity. Meeting summaries, research assistance and draft communications are sensible starting points because they address work that consumes valuable adviser time.
The adoption figures are increasingly substantial. Iress, citing Investment Trends’ 2026 Adviser Technology Needs Report, reports that 79% of advisers use AI in some form. The same article identifies inaccurate outputs and limited integration with existing systems as significant barriers. Those findings suggest that access and experimentation are moving ahead faster than consistently useful execution. Iress’s discussion of the research.
“In some form” leaves considerable room for interpretation.
An adviser occasionally drafting an email and a team embedding AI throughout preparation, discovery and follow-up might both count as users. Yet their commercial outcomes could be very different. Public adoption figures give us only a partial view of how deeply behaviour has changed and whether the investment is earning a return.
Even time savings require interpretation. An hour saved creates capacity. Management determines whether that capacity becomes better service, additional client conversations, lower costs or simply a less pressured working day. Each can have value, but the business case should be clear about which outcome it expects.
This is where established advisory practices deserve particular attention.
An experienced adviser with a substantial book has built a successful working method over many years. Changing it carries a perceived cost: disruption, learning time, concern about accuracy and possible damage to relationships they have worked hard to earn.
Some of these advisers will be older. Age alone, however, is a poor guide to willingness or ability. The more useful questions concern confidence, incentives, established habits and whether the new process demonstrably improves their practice.
A generic training session is unlikely to resolve those issues.
A capable practice manager can identify a specific frustration, introduce an approved tool into a real workflow and help the adviser judge the result. Starting with meeting preparation or follow-up may establish confidence. Involving the adviser’s support team can make adoption part of how the practice operates.
Winning the support of a respected senior adviser can also influence colleagues more effectively than another centrally issued presentation.
That is a management intervention, and it should be planned and resourced accordingly.
The early applications already show why the effort matters. Morgan Stanley’s Debrief was designed to capture meeting notes with client consent, identify actions, prepare a follow-up email for adviser review and save notes into its client management system. Citi reported that CitiScribe, its meeting summarisation tool, had been rolled out to all North American Wealth advisers in the first quarter of 2026. Morgan Stanley’s announcement, Citi’s announcement.
The immediate benefit is less administration. The more interesting possibility is better continuity between what a client says and what the institution subsequently does.
A concern about supporting an adult child, an approaching retirement or an unresolved estate issue can become a recorded action with an owner and a due date. An adviser can spend more of the meeting listening, then check the record and decide what requires attention.
Across the wider advice process, AI can assist with extracting information from documents, identifying missing evidence and preparing material for review. Providers are extending their offerings into advice drafting, quality checks and compliance support, although availability and maturity vary. Intelliflo, for example, distinguishes its live engagement assistant from capabilities still on its roadmap. Intelliflo’s published rollout information.
A product announcement does not establish widespread use or proven improvement in compliance. Nor does a polished file note establish that the underlying advice was sound. Human review must test the facts, reasoning and unresolved questions against the underlying evidence.
There is nevertheless a valuable connection between compliance and growth here. Better discovery can produce both a stronger advice record and a fuller understanding of the client. The same missing information that creates a documentation problem may conceal an important unmet need.
That brings me to an application I believe deserves much greater strategic attention: uncovering opportunities the institution has not yet recognised.
In my strategic work examining Citi’s wealth management opportunity, I explored how an institution could make more effective use of its existing relationships and information. The broader principle applies across the industry.
Relevant knowledge may be dispersed across meeting notes, banking relationships, investment holdings, client documents and appropriately licensed external sources. AI could help connect those fragments, surface a potential need, test the supporting evidence and prioritise it for an adviser’s attention.
Consider a hypothetical business owner. An earlier meeting recorded an intention to retire within several years. A subsequent public announcement suggests a possible business transaction. Other information indicates that the household’s succession planning may need updating.
A useful system would bring those signals together, identify what remains uncertain and suggest a timely conversation. It would give the adviser the evidence and context needed to decide whether to act.
The output should be a credible reason to speak with a client. The adviser still needs to establish what is actually happening and what the family wants.
Opportunity identification already exists in wealth management. Morgan Stanley’s Next Best Action is one established example. The strategic opportunity is to make discovery more comprehensive, better validated and more consistently acted upon. Morgan Stanley’s description of Next Best Action.
More alerts are not inherently useful. A stream of weak suggestions can consume the very capacity AI was meant to release. Management should measure which signals advisers accept, why they dismiss others and whether the resulting conversations benefit clients.
Commercial prioritisation also needs care. A model focused narrowly on near-term revenue could overlook an urgent client need with little immediate sales value. Client benefit, urgency, evidence quality and adviser capacity should all influence what receives attention.
The distinction between generative and agentic AI becomes relevant here.
Generative AI can produce a summary, draft or explanation. An agentic system can use models and connected tools to pursue a defined objective across several steps, adjusting its next action as it gathers information. The categories overlap: an agent often uses generative AI. Ordinary automated workflows can also handle many steps without needing an autonomous agent. Anthropic’s explanation of workflows and agents.
In wealth management, a generative tool might draft a meeting brief. An agentic system could investigate an approved signal, retrieve supporting records, identify gaps and prepare a proposed follow-up for review.
As its ability to act expands, management must define its authority precisely. Accessing information, changing a client record, contacting a client and implementing a recommendation involve very different responsibilities.
Larger projects require C-suite ownership and a clear place in the institution’s long-term strategy. Leaders need to understand both the capabilities they are funding and the responsibilities they are delegating. Data access, security, evidence trails, approval boundaries and accountability should be designed into the programme from the outset.
The Australian industry has had firm reminders of the importance of governance keeping pace with innovation. In its 2024 review of AI use among financial services and credit licensees, ASIC warned of a potential gap between accelerating adoption and the arrangements needed to oversee it. ASIC’s findings.
The same principle is explicit in the United States. FINRA’s 2026 Annual Regulatory Oversight Report reminds member firms that existing obligations—including supervision, communications, recordkeeping and fair dealing—continue to apply when they use generative AI. It also identifies particular risks from AI agents acting beyond their intended authority and highlights the importance of human oversight, traceable actions and controls on what agents can do. FINRA’s report.
For wealth management leaders, the implication is straightforward: delegating work to AI does not remove the institution’s responsibility for the outcome. Experienced managers have an important role in translating that responsibility into practical supervision as the technology becomes more capable.
The commercial discipline should be equally explicit.
Before a pilot starts, record the baseline, full costs and intended outcomes. Measure time saved after review and correction, changes in service quality, qualified opportunities, implementation rates and incremental revenue. Where practical, compare results with similar teams outside the pilot, accounting for differences in their books and market conditions.
Licences are only part of the cost. Integration, data preparation, training, supervision and management time also count. Avoid treating released capacity as both a cash saving and a revenue gain unless both have actually occurred.
Middle managers are particularly well placed to connect these measures with daily practice. They can see whether a workflow is being used, whether it creates rework and whether a promising opportunity stalls at the handoff between teams.
They need authority and capacity to do that work. Making them responsible for adoption while leaving their workload, incentives and decision rights unchanged is a weak implementation strategy.
Ultimately, the human relationship remains central to my view of wealth management.
In my experience, clients want someone they can talk to about their finances and family affairs. They also value a human connection to the institution backing the advice. They want to know who will take responsibility, coordinate specialists and remain available when circumstances change.
That gives institutional wealth management a potential strategic advantage: a trusted adviser supported by the knowledge, resources and continuity of a larger organisation. The advantage has to be earned through the client’s experience.
AI can help that institution remember more, recognise needs earlier and follow through more reliably. Experienced leaders must ensure those capabilities strengthen the relationship.
This is the intersection that interests me most: applying practical wealth management experience to decisions about AI, then helping advisers and teams use it effectively.
Senior management needs enough understanding of AI to make sound strategic choices. Practice leaders need the credibility and skill to translate those choices into daily behaviour. Both need evidence that the changes are working.
The return will be earned when a better-informed adviser has a useful conversation, a client receives appropriate help and the institution delivers on its commitments.