The success of AI in wealth management may depend less on which tools firms buy than on how well they prepare their businesses to use them.

That is the view of Kevin Andrews (main picture), chief operating officer at , who has contributed to WealthBriefing’s forthcoming AI in Wealth Management research report. His responses highlight a shift from exploring the technology’s capabilities to establishing how it can deliver practical value within a regulated business.

“12 months ago, we were largely exploring what AI could do. Currently we are focused more on how it can be used safely and practically within the business,” Andrews said.

Bentley Reid has introduced an AI policy and governance on its use, with staff using tools such as Copilot for research, summarization, drafting and everyday productivity tasks. The firm does not allow AI to make investment decisions or generate client advice.

For Andrews, the most significant change has been moving the discussion towards the business processes and data that support the technology.

Creating capacity for client service
Asked where AI could make the biggest difference to the economics of wealth management, Andrews points to productivity.

Relationship managers spend considerable time preparing for meetings, writing reports, documenting conversations and finding information. Reducing that workload could give them more time for activities that benefit clients.

“As a boutique wealth manager, we remain a people business. The real opportunity for us is helping our staff serve clients more effectively and consistently, while creating additional capacity as the business grows.”

However, demonstrating a promising use case is only the first step. Turning it into something reliable, controlled and supported within daily operations requires a stronger business case.

“It’s largely about proving that the value is real and repeatable,” Andrews said.

A useful test is whether staff are still using a tool six months after its introduction because it continues to make their jobs easier.

Adoption depends on everyday usefulness
Some of the most effective applications Bentley Reid has found so far are straightforward: summarizing information, preparing first drafts, extracting meeting actions and helping structure documents.

“They aren't particularly glamorous, but they save time and solve real problems,” Andrews said.

Training helps staff understand both the capabilities and limitations of AI. But adoption ultimately depends on a tangible benefit. A tool that saves time has a clearer route into everyday use than one that adds another login or process.

Data presents another challenge. Bentley Reid has a mixture of established and newer systems, with information held across different platforms. Making that information structured, accessible and understood is a priority alongside identifying practical applications.

Human accountability remains central
Bentley Reid’s measured approach also reflects the need to balance opportunities with appropriate controls.

“Our starting point is that responsibility remains with the individual, not the AI,” Andrews said.

Generative AI can produce convincing answers that are incorrect. At Bentley Reid, someone therefore remains accountable for reviewing and approving the final output, with that principle reinforced through policy and training.

Looking ahead, Andrews expects the strongest adopters to be firms that establish the right foundations and integrate AI effectively into their operating models. Poorly understood data and inconsistent processes risk being amplified by the technology.

His perspective offers a practical question for the wider industry: can an AI initiative deliver a lasting improvement in how people work?

This article draws on an interview conducted for WealthBriefing’s AI in Wealth Management research report, due for publication later this year. Private banking and wealth management professionals are invited to contribute their views through the accompanying survey.