The editor looks at clients' use of AI tools to research financial material and even suggest ideas for investment, financial plans and more. What are the benefits and risks? Figures from law, technology and wealth management explain the territory and what lies ahead.

(An earlier version of this analysis appeared on Family Wealth Report, sister news service to this one. The implications of these issues are global, so we hope readers across the financial jurisdictions of the world find these comments useful.)

AI resembles a “five-year-old genius” that can cut through information and come up with answers faster than the sharpest student. A question, however, is how far HNW individuals use it to plan their finances.

As of today, while the technology has grown rapidly, it appears that while AI can assist with basic tasks, it does not yet have the personal touch or the accountability and regulatory oversight that a financial advisor should have.

Surfing the internet, your correspondent found examples of YouTube videos about topics such as “Asking ChatGPT for financial advice,” “Get Out of Debt Using ChatGPT,” and "How To Use ChatGPT-5 to Build INSANE Financial Models.” A Reddit page has the headline: “I use copilot as a financial advisor, is that a good idea or should I make an appointment with a real one (that is if I can afford it)?”

It’s clear that this is a trend. But one of the issues that came to the surface following the 2008 financial crisis was that advisors needed to be more mindful of their fiduciary responsibility to clients. If that is important, where does AI fit in, given that it may have biases acquired from from underlying data?

The soft power of seduction
As use cases have proliferated, one concern that has arisen is not that AI puts humanity under tech overlords, but something softer and more insidious. The University of Tennessee law professor and writer, Glenn Reynolds, has caught the mood with a new book, Seductive AI. As the online description of his short book says, “It [AI] is not about domination, but seduction. There is no doubt that AI is useful, but `being useful is the subtlest form of seduction there is’.” Central to professor Reynolds’ contention is that AI has an embedded desire to be liked. And that’s dangerous, he writes.

A reason why this matters for wealth managers and advisors is that if the concept “trusted advisor” has any meaning, it is about telling clients what they might not want to hear and asking insightful questions. Being able to say “no” at times and act as an honest sounding board is vital. A question is whether AI can ever do that consistently.

Mainstream adoption
“AI started to become a consistent conversation topic for our clients around 2023 – and the urgency accelerated sharply after generative AI tools started seeing mainstream adoption,” Warren Finkel, managing director, Omega Systems, a US firm, told this publication in a recent interview. He is generally positive about aspects of AI but has caveats.

“What we hear consistently from financial services firms is that younger investors are arriving to conversations better informed – but not necessarily more accurately informed,” Finkel said. “They've done their research, often using AI tools, and they have real opinions. That's actually a positive shift in engagement. The challenge is that AI outputs can look authoritative without being vetted, contextually appropriate, or current. A model trained on historical data and general principles isn't calibrated to an individual's specific situation or the regulatory environment their advisor operates in. The firms that handle this well are the ones treating AI-influenced clients as more engaged, not more difficult.”

Eduardo Arista, a US-based partner at law firm Holland & Knight, said AI is "increasingly visible" in client conversations. He gave examples such as questions sent in chatbot form; jurisdictional comparison tables; a model produced for them, and second-opinion requests after a client has already run the topic through a model. "A chatbot will not replace a trusted advisor. What erodes the relationship is the advisor who cannot engage with the AI material a client now routinely brings into the room. When a client hands you a chatbot memo, they are asking, implicitly, whether you can do more with it than they did. The advisor who cannot answer that question on the spot has already lost ground," he said. 
 
This news service asked Arista whether clients talk about their own use of AI with their advisors.

"Whether they talk about it depends on the client. Sophisticated principals discuss it openly. Some hide it because they are testing whether the advisor will reach the same conclusion. And the next generation in UHNW families, the 35 and 45 year-olds who will eventually inherit, increasingly bring AI output into the room as a way of pressure-testing the lawyers their parents have used for years," he said.

Arista listed the following examples of what tasks clients set via AI: Structuring options across jurisdictions; side-by-side comparisons of where to hold assets or where to relocate, the kind of table a lawyer would not produce in a single sitting; tax outcome modelling for events such as a move, an exit, a gift, or a death; second opinions on what an existing advisor has already recommended; and drafting work – trust language and letters of wishes.

Pros and cons
Joshua Landsman, wealth strategist and senior vice president at Wilmington Trust, based in Palm Beach, Florida, sees risks and opportunities for clients and advisors alike.

“People are risking not getting tailored advice to [meet] their needs because they are not providing all of the relevant information to the AI system being used to get the appropriate advice,” he said. “AI is best used when the prompts are very specific. Clients typically are unable to identify all the potential issues based on their specific facts because of a lack of knowledge, experience and awareness of those particular issues.”

AI use is now mainstream. According to the 2026 Global AI In Finance Report from KPMG, more than three-quarters of organisations use AI in financial planning, reporting and commercial analysis. 

Almost every day, a financial services firm announces a new AI-powered offering. Earlier in May, US brokerage giant Charles Schwab said that its first generative AI capability for retail clients had been made available. The firm said one of its recent surveys showed that among nearly 1,000 retail clients, more than 60 per cent are interested in using AI, and nearly 70 per cent think it can either support routine tasks or play a meaningful role in investing when paired with human expertise.

Tools in the box
Wilmington Trust’s Landsman said clients indicate that they use ChatGPT, Gemini and social media to research planning concepts. 

“For example, in one conversation with clients, they indicated that they wanted to consider a particular trust structure based on their research using AI.  The clients were researching trust planning strategies to leverage before the sale of a business and AI mentioned a Charitable Remainder Trust as an option. 

“I asked the clients how much they gave to charity on an annual basis and they said less than $25,000. I explained that while a CRT would help them defer capital gains taxes, ultimately, assets would pass to charity and, if they weren’t charitably inclined, other strategies may be a better fit for them. We reviewed their cash flow needs as well as legal documents and suggested they consider a pre-transaction sale to an intentionally defective grantor trust. Therefore, while AI identified a potential solution for the clients, AI wasn’t able to have a deeper discussion because the clients didn’t know how to evaluate their own situation. I explained why the trust structure did not apply to them based on their goals and objectives,” he said. 

“People are risking not getting tailored advice to [meet] their needs because they are not providing all the relevant information to the AI system being used to get the appropriate advice. AI is best used when the prompts are very specific. Clients typically are unable to identify all the potential issues based on their specific facts because of a lack of knowledge, experience and awareness of those particular issues,” he said. 

There is also an age difference to consider.

“I see more enthusiasm because many younger clients are not used to having an advisory team to assist them and some parents do not introduce their advisors to their children until much later in life. Because of this, I see younger people leaning on AI more because that is all they know from an advice perspective. I have not seen a huge adoption of AI use by older clients, however, when used, it’s used to do initial research on a topic with a request for an explanation by us as advisors,” Landsman said. 

Even if clients don’t use AI greatly, it is important for them that their advisors are fluent with it, he said. 

“AI is a productivity tool that can help clients gather data about their financial lives and then provide that organised data to advisors to help them make decisions. I see AI being effective to help clients create balance sheets and cash flow analysis for example.

“For clients, a professional advisor that has experience navigating clients with varying levels of net worths through different market cycles should remain invaluable as opposed to AI which requires a great level of specificity and direction from the user in order to optimise its value. Advisors can use AI as tools to gather information about their clients, organise that information and translate it into outputs to help clients make financial planning decisions about retirement, making a big purchase, planning for inheritance, etc,” Landsman said. 

Not going away
In any event, AI is part of today’s reality, Omega Systems’ Finkel said.

“AI-generated financial guidance isn't going away – and advisors who dismiss it outright will lose the room. The better approach is to engage with it. Understand what your clients are reading, ask questions about where they're getting their information, and be ready to contextualise it rather than simply contradict it,” Finkel said. “AI tools can surface ideas that are interesting starting points, but they don't know a client's full financial picture, their risk tolerance, their tax situation, or their long-term goals. The advisor's job is to be the layer of judgment that AI can't replicate. Younger investors in particular respond well to advisors who take their curiosity seriously and can translate complex concepts into plain language – which is exactly what a good advisor should already be doing.”

Gaps and shadows
Arista argues that there is still a space for human advisors, and they need to remind clients about it. 

"The model is good at producing something that looks like the answer. Whether it is actually the answer is the question the client is no longer equipped to evaluate on their own. That gap is where the advisor still earns the fee. The move that works is curiosity. If a client brings an AI memo to a meeting, read it with them. Show them where the model is right, where it is wrong, and why the gap matters. The result is not a defeated client. It is a client who has watched their advisor demonstrate, in real time, what 30 years of training adds to a tool they can run themselves," he said. "Telling clients not to use AI is a wasted breath. They will use it. The advisor's job is to teach them when to trust the output and when to bring it to counsel before acting on it."

Finkel pointed to some of the challenges AI brings advisors and by extension, their clients.

“AI has accelerated the information gap between what financial firms' investors think they know and what the firms themselves can verify or stand behind. That creates a real challenge for advisors – and an equally real risk surface for the firms we work with. The volume of AI tool adoption across financial services has created a significant shadow IT problem. 

“Firms are deploying tools faster than their governance structures can absorb them – employees are using AI to draft communications, analyse documents, and interact with financial platforms, often with no policy, no data inventory, and no clear understanding of where that data goes. Every one of those use cases is a potential security or compliance exposure,” he added.

Guardrails
Arista said there are four "guardrails" that advisors and their clients should consider. Firstly, have an an approved-use policy that names which tools are sanctioned, for what purposes, and under what conditions, he said. "Open-ended use across every tool available is the most common source of trouble," Arista said. Second is data discipline: "Privileged information, sensitive financial detail, and personally identifying material do not go into a consumer tool that uses inputs for training. Once that happens, the privilege is at risk."

Third and fourth are a need for human review of material decisions, and vendor diligence on the tools themselves. "Anything that creates a tax position, transfers wealth, or modifies a structure goes through a qualified advisor before execution. AI output is a starting point, not a substitute. The same scrutiny [that] family offices apply to custodians and trust companies belongs on AI vendors," he added.

A growing force
The financial planning sector certainly appears to see AI as a major force.

In an article from the May 2025 edition of the Journal of Financial Planning, the author, Emma Foulkes, wrote: “From a consumer’s perspective, we see increasing numbers relying on AI to take material decisions on their behalf, mediate their interactions with financial markets, and finally automate their financial lives.

“Right now, AI is mostly used as an assistive tool to explain concepts and options. Others already use them as advisory systems that recommend actions. 

“When implemented well, AI can enable firms to innovate at pace and better meet customer needs. It can enhance good outcomes by improving personalisation, improve customer understanding and support and driving better quality services at lower cost. Agentic AI in particular could support people to automatically optimise their household finances, reducing inertia through automatic switching and potentially encouraging a saving and retail investment culture, which could help us respond to demographic changes in the UK.”

Foulkes warned that AI can amplify risks, such as embedded biases, discrimination, exclusion, opaque decision-making (particularly when multiple AI models interact), misleading or hallucinatory advice and eroding consumer trust. “It could also introduce new risks if decisions are increasingly delegated to AI agents, including reduced consumer agency, reduced consumer understanding, unconscious manipulation and further decrease financial literacy,” she wrote.

Back in 2013 in the UK, new advisor regulations called the Retail Distribution Review forced some sales-based IFAs out of business, creating what some feared would be an advice gap. Into that “gap” moved digitally-driven wealth platforms, or “robo-advisors,” with the likes of Nutmeg, as it was called then, being the most well known. (The Nutmeg brand disappeared as the firm was eventually absorbed into JP Morgan.) A number of other “robos” took flight, for example in the US with Wealthfront and Betterment. While “robo-advisor” isn’t much used as a term these days, it is arguable that AI represents a new stage of it.