T. Rowe Price: Limit AI messages to routine tasks

T. Rowe Price warns in a new practice management report that financial advisors who rely on artificial intelligence to write client messages risk eroding client trust and should confine automation to routine, transactional communications.

The firm examined client responses when messages felt machine-generated and found that unpersonalized, generic AI-written communications can prompt distrust and harm long-term relationships. The report identifies service confirmations and scheduling as examples of messages that clients generally accept when automated, while emotional topics perform poorly when an advisor’s voice is replaced.

Emotional topics the report highlights include condolence notes, congratulations and messages tied to personal milestones. The report notes that a single impersonal email or an overly formal greeting can reduce the familiarity advisors have built with clients over years.

To explain the effect, the report references behavioral research showing people remember negative interactions more strongly than positive ones. It cites a survey finding that 71% of U.S. adults would view a candidate’s speech less favorably if they learned AI helped write it, compared with 38% who felt the same about a pop song. A study cited in the report also found that heavy use of automated messages for emotional subjects can provoke what researchers describe as “moral disgust” in recipients.

Small personal details matter, the report adds. It uses the example of automated systems addressing a client as “William” when the client prefers “Bill.” Such mismatches do not usually end relationships on their own, but they can chip away at familiarity over time. The report recommends advisors avoid applying their signature to a fully AI-written note and suggests using time saved by automation for phone calls or handwritten messages.

The report offers a practical test: would this message feel personal if I were the client? If the answer is no, the report recommends reconsidering automation. The guidance comes as financial firms increase experiments with generative AI and focuses on keeping authenticity, transparency and human judgment in client-facing communications while permitting automation where it poses limited relationship risk.