Russell Hutchinson discusses how advisers can use artificial intelligence into their practices.

Artificial intelligence is moving rapidly into financial services. Tasks that once appeared to require human expertise can now be handled, at least superficially, by an online system.

The question for advisers is not simply how far artificial intelligence will reach. It is also whether the advice profession can define and demonstrate value in areas where automation remains weak.

A hierarchy of financial advice activities helps explain the challenge. The lower levels – such as accessing prices, product information and basic comparisons – have been contested by online services for some time. Yet after many years in the market, online services remain relatively niche, albeit growing.

Providing information alone does not appear to capture a large share of the market. AI may change that because it can engage in a conversation that at least sounds a lot like advice.

AI is, of course, already giving advice. We periodically test AI systems by asking them for financial advice. Until a few months ago, the usual response was a polite suggestion to consult a financial adviser.

That is changing.

In a recent test, the system asked for more information, personalised its response, and then made specific recommendations about product providers, types of cover and premium structures.

When these connect to agents, they will move to execute purchases as well. That’s worrying, because there were significant problems with the advice.

It referred to product types and premium options that are not available in New Zealand. It recommended specific providers without properly substantiating the reasons. Its premium estimates appeared to rely on out-of-date information taken from an adviser’s website. Its product comparisons were wrong.

In our view, the advice also raised serious questions about compliance with the Code of Professional Conduct and the legal restrictions on providing regulated financial advice without the appropriate licence. Our regulator has not yet acted, perhaps because there is not yet a link to purchasing.

The important point is not that AI is currently unreliable, but that it will be trusted by people to help them execute.

Advisers are likely to meet clients who have already asked AI about their insurance or investments. Some may arrive with useful questions. Others may arrive with a confident but inaccurate set of conclusions.

The digital opportunity

The conversational nature of AI could nevertheless make online services considerably more useful. For advisers, the most obvious response is to use the technology themselves.

Call recording, transcription, summarisation, document processing and system integration can reduce administration and improve consistency. These benefits do not necessarily require a radical change to the advice proposition.

But the advice proposition may need some thought in a world where many clients are trusting AI to enable them to bypass your services.

The advisers best placed to navigate this change may be those who adopt a behavioural approach that seeks to add value throughout the client relationship.

Many adviser systems have been designed around feature improvements: faster data entry, better document production or more efficient research. Those are worthwhile. But use-case improvements may be more significant. Some of the most effective work on the system with the client sitting beside them.

They compare options, test assumptions and discuss the implications together: premium comparison, product research, underwriting information and relevant social proof can all become visible evidence of the process. That can make trust easier to establish.

You are not asking the client to rely solely on personal confidence or reputation. The client can see how the recommendation has been developed and how it relates to the broader market.

Research in financial planning points in the same direction. Studies reported higher satisfaction where advisers considered the broader marketplace and presented a range of options. Push this idea further and the model begins to resemble coaching.

An accountant, nutritionist, piano teacher, or gym instructor typically reviews evidence, compares it with a goal, identifies a gap and agrees on actions before the next meeting. The relationship is not simply about delivering a product. It is about helping the client improve their position over time.

AI can help you integrate the information from a range of tools - creating more speed and much nicer, more personalised documentation. You can provide judgement, accountability, context and encouragement. The combination offers a level of engagement that neither technology nor human expertise is likely to deliver alone.

Research from Zurich and OnePath Life in Australia, available through The Client Connection, explores the importance of this ongoing relationship. It is worth considering as advisers assess not only how their services are delivered, but what clients are really being helped to achieve.