Generative AI (GenAI) is entering financial advice workflows, promoted to reduce administrative load and free advisers to spend more time with clients. Behavioural research by Morningstar puts the client at the centre of the adoption question: investor reactions vary depending on what AI is used for, how it is used, and why the adviser appears to be using it.
That framing matters because the perceived value of advice is shifting. Morningstar’s report describes a move from “investments” to “investors”, with clients expecting more coaching, guidance, and reassurance – the “softer side” of advice – while time remains the binding constraint. In that context, GenAI can look like a practical tool or a threatening signal, depending on how it shows up in the client experience.
These insights are contained in Morningstar’s “What do investors think about generative AI In financial advisors’ workflow?” report (September 2024).
In the study, Morningstar asked 1 218 investors to consider hypothetical scenarios in which their adviser uses GenAI for different use cases. Investors then answered follow-up questions measuring comfort, relationship impact, and willingness to pay, alongside additional questions used to explore what was shaping those reactions.
A separate group of investors (636) considered the same use cases with the adviser completing the task without any mention of GenAI, providing a baseline to assess how introducing GenAI changes investor reactions.
The key result: reactions are use-case specific
Morningstar tested six specific use cases, ranging from marketing and summarisation to personalised emails and recommendations. Across the study, investors were generally more comfortable when advisers completed tasks without AI, but they did not report broad discomfort with AI across all uses.
Where the differences emerge most clearly is in relationship impact – the aggregated measure covering whether the adviser is acting in the client’s best interest, understands the client as a person, and will help the client to reach financial goals. On that measure, marketing, summarisation, and research support produced less negative relationship impact than personalised emails and personalised recommendations.
The sharpest case is AI-written “personal” communication. Morningstar describes a “large negative effect” when AI was used to generate a personalised email, noting that investors appeared to react viscerally to the idea of an adviser trying to replace “a human interaction and a display of companionable intimacy” with AI.
Because comfort and relationship impact are separate measures, the report allows for scenarios where investors are not highly uncomfortable with AI in principle but still revise relationship perceptions downwards in specific contexts.
At the same time, Morningstar’s researchers caution against turning that pattern into a rigid rule. Senior behavioural insights researcher Samantha Lamas said: “Our research shows that there isn’t a simple bright line where AI suddenly becomes a ‘relationship risk’, which makes sense given the complexity of social relationships and the evolving use of GenAI.”
Instead, she said the study provides a way to pressure-test any use case – one that fits real workflows better than a “low versus high intimacy” label.
“Our research does provide a way to pressure-test a GenAI use case by the adviser asking themselves three questions: what are you using AI for, how are you using it, and why are you using it?”
She defined each element: “The ‘what’ is about human substitution: Are you automating something that shapes trust, like communication or advice framing? The ‘how’ is about safeguards: Are you being transparent, protecting client data, and ensuring accuracy? And the ‘why’ is about client benefit: Is this improving the client experience, or just saving the adviser time?”
Lamas added a qualifier: the “risks associated with using GenAI don’t come from intimacy alone. It comes from misaligned uses that compromise accuracy, obscure the personal relationship, or prioritise efficiency over client outcomes”.
Willingness to pay
The study’s willingness-to-pay results capture a different kind of judgement: what clients think advice is “worth” when AI enters the picture.
Morningstar reports that investors consistently set lower pay expectations for an adviser using GenAI than for an adviser working alone – with drafting marketing materials the exception.
The report frames this as a consumer inference problem: people associate pay with skill and effort, so AI can read as “less work”, even when it improves efficiency. That inference is also visible in the themes Morningstar extracted from investor explanations, where “efficiency” can be positive but can also be interpreted as cutting corners or lowering quality.
Asked whether Morningstar believes the adoption of GenAI will create structural pressure on fees, Danielle Labotka, behavioural scientist (saving and retirement), said there are two theoretical directions. Advisers could decide to discount the fees on services rendered with the assistance of GenAI (which “would be most in line with what we found”). On the other hand, advisers could decide they want to focus on raising the cost on those human-specific services for which they do not employ GenAI. She added: “As of yet, I haven’t heard of anyone taking either of these approaches.”
Her expectation, at least at this stage, is that firms will treat AI less as a pricing lever and more as a capacity lever.
“The approach advisers are taking is not to fiddle with their pricing structure but rather to focus on providing more of those high-value human services (for example, behavioural coaching, life planning, etc) to their clients with the time they save on other tasks using generative AI.”
Labotka expects more pricing pressure only when AI becomes more thoroughly embedded in different parts of financial planning, while “at these earlier stages, I don’t think the technology is being deployed deeply and consistently enough to start applying pressure directly to pricing”.
What investors want before they relax
When Morningstar asked investors what would make them more comfortable with advisers using GenAI, five themes dominated: protection of data (35%), transparency in usage (33%), human oversight (27%), preserving client agency (15%), and unbiased advice (8%).
The distribution points to practical, client-facing concerns – privacy, clarity about use, and human accountability – alongside a smaller concern about bias.
In response to questions from Moonstone, Morningstar’s researchers translated two of those high-priority themes – transparency and human oversight – into guidance on what disclosure and review should look like in the client experience.
On disclosure, Labotka acknowledged the concern that bringing up GenAI could provoke client anxiety but said other research indicates that the disclosure risk advisers face with GenAI is not more heightened emotions but potential client disengagement. Advisers do not want disclosure of GenAI to cause their clients to disengage, but the failure to do so can erode trust.
One way for advisers to be transparent with their clients and prevent disengagement is to be clear when the output has been assisted by AI rather than solely originating from it.
Labotka said “good disclosure doesn’t have to be complex” and should use “clear, concise language that is exact and not overly apologetic”. As an example, a disclosure line on a client guide can read: “Our staff used generative AI to create the outline for this guide, and then, we wrote the content.”
On oversight, Lamas said: “For clients, credible oversight is less about process and more about accountability and clarity. Most investors do not need a very detailed explanation of how AI is used, but they do need confidence that a human is using their own judgement throughout the process and ultimately responsible for the outcome.”
She added that “it is often enough” to explain that AI is used as a support tool “for things like research or synthesis”, while “all final decisions and recommendations are made by a human adviser”.
Oversight really becomes credible when the adviser can “stand behind the work” and “clearly explain their recommendations: what factors they considered, what alternatives they evaluated, and why one option was chosen over another”.
Making efficiency persuasive
Morningstar reports that investors who perceived GenAI use as efficient were more likely to respond positively across the use cases.
Lamas said advisers should translate efficiency gains into tangible improvements clients can observe – faster follow-ups, more personalised communication, fewer errors, or more attentive meetings – and make the benefit explicit at the moment the tool is used. For example: “I’m using AI to capture notes during our conversation so I can stay fully focused on you and make sure nothing gets missed.”
The report also identifies wider factors that shape how easily clients accept that efficiency framing. The appendix notes that belief in AI’s accuracy and a general trust in AI were associated with more positive responses across comfort, relationship impact, and willingness to pay, while privacy concerns reduced comfort in the personalised recommendations scenario.
Practical actions advisers can apply
To help advisers implement AI in a way that strengthens client relationships, Morningstar provided a summary checklist built around five the focus areas mentioned above – data protection of data, transparency, human oversight, client agency, and ensuring AI objectivity.
Practical actions from the checklist include:
- Avoid automating human interaction. Clients react negatively when GenAI is used to replace human interaction. Where possible, use it to facilitate better conversations rather than to substitute for them.
- Address privacy concerns clearly. Be upfront about privacy policies, how client data is protected, and what safeguards are in place.
- Use transparency and choice. Develop disclosure statements explaining how you use GenAI and consider giving clients an option to opt out of GenAI use for their account.
- Treat GenAI output as a “first draft”. Given GenAI’s tendency towards hallucination, build a documented review process to quality-check outputs for accuracy.
- Communicate client benefit. Explain that AI helps you do certain tasks more efficiently, so you can spend more time on the parts of advice clients value – context, judgement, behavioural coaching, and reassurance.
AI tipping point?
Investor expectations around AI are likely to evolve as adoption becomes widespread. Does Morningstar foresee a time when not using AI signals that an adviser is “out of date”?
“For me, it’s hard to envision a point when advisers who don’t use generative AI are seen as out of date, if only because advising is increasingly a human business,” said Labotka. “However, you can ask yourself whether someone would’ve said the same thing about using a computer a couple decades ago.”
If a tipping point does come when advisers must use GenAI or lose their business, she said “it won’t sneak up on them” but will be the result of “sustained, increased trust in the technology”.
Labotka said it is important that advisers who want to “stay ahead on GenAI” do so in a way that does not erode trust.
“This starts by identifying promising uses of generative AI. If an adviser wants to try generative AI for a new task, we recommend asking themselves the following questions: Is this a task I would benefit from offloading to generative AI? Is this a task generative AI does well? Will my clients be comfortable with generative AI being used for this task? If the answer to all those questions is yes, then you have a promising use of generative AI.”
