Financial advice firms adopting artificial intelligence must invest as heavily in advisers’ judgement, communication and other soft skills as they do in the technology, according to David Tait, founder and managing director of Redmill Advance.

Tait made the comments following the launch of Claude for Financial Advisors, a platform designed to support research and administration, which Redmill Advance described as another sign that AI is moving from experimentation into the everyday workflows of financial advice.

He said the bigger question was not how much AI would automate, but what capabilities advisers would need once it took on more of the administrative side of the role.

“Tools such as Claude can connect directly to CRMs and portfolio platforms, helping advisers spend less time on administration and prep,” Tait said.

“But as AI increasingly handles routine tasks, the skills that differentiate advisers will shift further towards good judgement, great communication and the ability to challenge an AI-generated output rather than simply accept it. Firms cannot assume that introducing sophisticated technology automatically makes their advisers more capable.”

Tait said firms should resist adopting tools before addressing governance gaps internally.

“Firms need to ask whether their data is accurate, whether advisers truly understand how to question what an AI produces, and just as critically whether the technology aligns with UK regulatory expectations, such as Consumer Duty,” he said.

The Financial Conduct Authority’s Consumer Duty sets standards of consumer protection across financial services and requires firms to put their customers’ needs first.

‘AI is only as reliable as the information it can access,” Tait said. “Incomplete or inconsistent data can produce convincing outputs that miss important context.”

He argued that one of the most significant risks was advisers becoming overly reliant on outputs that are only partially informed.

“A fluent AI assistant can generate a highly convincing answer from incomplete or inconsistent data and still present that answer in a way that feels competent,” he said. “Hallucination is an obvious risk, but an unchecked AI output can be incomplete, missing key information relevant to a client.”

“In a regulated and relationship-driven sector, what appears to be a confident answer from the AI but is built on partial information or incomplete client data can become a costly mistake for the client and the adviser.”

Legal experts have separately warned that poorly governed AI systems expose UK businesses to legal, financial and reputational risks, including from inaccurate outputs.

Tait said advisers needed detailed technical understanding to recognise when an AI output should not be relied on, and remained responsible for turning information into advice that reflects a client’s circumstances.

“AI can accelerate the process, but advisers still need the competence to evaluate the output, understand its limitations and take responsibility for the advice ultimately given to the client,” he said.

According to Tait, human advisers hold tacit knowledge about client relationships, including a client’s history, family circumstances that can affect future planning needs, and their wants and desires.

Tait said the rise of adviser-specific AI strengthened the case for a dual approach: adopting the technology while investing in continuous professional development.

“Businesses must invest in people as deliberately as they invest in technology,” he said. “That means understanding specific adviser development needs, embedding effective training programmes that meet those needs, and ensuring advisers continue to develop the skills needed to exercise sound judgement.”

He said the same applied to the next generation of advisers, who would need ongoing learning integrated into their job roles.

“If AI is used to substitute adviser knowledge, firms risk creating advisers who can review an output but lack the underlying understanding to recognise when something is wrong,” he said.

On whether AI will replace human advisers, Tait said: “In my view, the answer is probably not. But it will change the skills that advisers need to work more effectively in tandem with these emerging AI technologies.”

“Technical knowledge will remain important, but so will great communication, empathy, and the soft skills that connect the adviser to the client emotionally,” he said. “Clients are paying for trusted advice when decisions are complex. Human capabilities will only become more valuable as AI continues to advance.”

Read the full article →