The robots are coming and they are after our jobs, so we’ve been told.

Huge advances in artificial intelligence mean machines are now capable of doing many of the tasks once carried out by humans.

That’s good news for employers, who will benefit from greater efficiencies, but it could potentially mean bad news for less skilled workers.

By the mid-2030s as many as one in three jobs – both skilled and unskilled roles – could be lost to automation, according to consultancy PwC.

While manual positions such as factory workers and waiters are most at risk, automation does present a credible threat to more highly-skilled professions – including financial advisers.

And in a way, that threat is already very real, with robo-propositions making some inroads into parts of the advice process in the past few years, albeit we’re starting to see some high-profile failures here too.

Some believe that threat will only get greater as AI improves. After all, computers are better than humans at analysing large sets of data, can carry out tasks much quicker and make data-centric decisions, rather than on human instinct and experience, which is often fallible.

Nevertheless, we believe technology has a long way to go before it replaces human advisers.

Perhaps the biggest barrier to widespread use of AI is regulation. Trying to explain why a machine learning model has behaved in a certain way is difficult, even if we are confident it has produced the right outcome.

Therefore, we are asking the customer to trust the outcome it produces. It is highly unlikely that will go down well with the Financial Conduct Authority, which will no doubt be keeping a close eye on the adoption of AI in the financial advice industry.

Then there is the question of data – something you need a lot of for a successful AI model. All but the biggest advice firms have the sort of data needed to create successful AI models.

Platforms could aid smaller advice firms, by using AI to ‘nudge’ advisers to review certain portfolios, to help them stay on top of things and ensure better customer outcomes. Of course, this comes with its own regulatory challenges, namely: whose data is it, and could these nudges amount to a personal recommendation?

Moreover, data is by no means perfect and can contain biases. An AI model’s behaviour is defined by the data that feeds it. If that data set contains biases, then it will taint the model and there is the potential for bad client outcomes.

On a more basic level, people are typically suspicious of new technology, which can slow down adoption. It is questionable whether financial advice firms will be comfortable basing a recommendation on algorithms they don’t fully understand themselves.

That suspicion also stretches to the general public. While younger, less-wealthy customers may be willing to interact with a chatbot or a robo-adviser, older clients may be less willing to do so.

And even if attitudes change, there is a very good chance that people – young or old – will still want the peace of mind of speaking to a human when it comes to more complex transactions.

Therefore, we believe that rather than threaten advisers’ existence, AI will instead be used to better equip advisers to provide quality advice, potentially by using it to improve operational efficiency and gather better information about their clients.

For example, machine learning could help advisers to understand and predict customer behaviour, such as how they are likely to react in a market downturn. This would allow the adviser to intervene before clients take steps that could lead to bad outcomes.

Machines learning’s ability to categorise indicators and predict likely behaviours means it lends itself well to risk-profiling. As well as a client questionnaire, there might also be demographic information, such as their age and where they live, that is relevant.

So, while we will undoubtedly see AI used more and more in the provision of advice, we believe it will be used primarily in speeding up information collection and to automate tasks that take humans several hours to complete.

The robots are coming. But they’re not coming for financial advisers – not yet.

This article originally appeared on Money Marketing on 16th July 2019.

About the author

Simon Clare

Simon Clare

Global Chief Technology Officer

Based in London, Simon is the Global Chief Technology Officer for Bravura Solutions. He is responsible for managing Bravura’s technical product roadmap and identifying how emerging technologies impact the wealth management and savings industry, particularly the technical and ethical questions surrounding the use of machine learning in financial services. Simon has an academic background in software engineering and mathematics. His career has primarily been spent in financial services, with experience in software development, solution architecture, product management and strategy. He has been with Bravura since May 2013.

More Insights