As debates around the relative merits of automated and face-to-face advice continue to rage, onlookers are assessing whether more advanced artificial intelligence systems that learn and adapt by themselves could radically shake up the planning profession.
In October 2017, the government released an independent review on the increasing benefits AI could have for the UK economy.
The report concludes that: “While human financial advice is costly and time-consuming, AI developments, such as robo-advice, have made it possible to develop customised investment solutions for mass market consumers in ways that would, until recently, only have been available to high net worth clients.”
Other countries have made headway with AI solutions, but their uptake across advice, wealth management and pensions in the UK has been lukewarm, despite government estimates that AI could add £630bn to the economy in 20 years.
Money Marketing spoke with advisers, consultants and technology providers to see if AI can integrate effectively with human advice to give clients optimal outcomes.
The government’s 2017 Budget allocated £75m to developing intelligent machines within the UK, and to establishing the Centre for Data Ethics and Innovation to conduct research. The government’s review confirmed personalised financial planning, fraud detection, and anti-money laundering are the three areas with the most AI potential.
For financial planners, crucial process automations could include both customer-facing services and back-office functions.
Owing to the difficulties of transitioning staff into new and more complex roles, Altus Consulting, which recently released a white paper on advice AI, says financial planning firms must remain agile and adaptable. The white paper says: “The successful firms in 10 years’ time will have made more right decisions than wrong and recognise that AI can help make them faster and better. The important step is to make decisions, continually learn, check back, and go again.”
Aside from eliminating human bias in investing, better use of AI may cut costs for financial advisers increasingly listing time management and compliance costs as their main concerns.
How AI crept into investment
|1956||“Artificial intelligence” is first coined as a term in the USA|
|1969||First international AI conference is held at Stamford University|
|1973||UK parliament commissions Sir James Lighthill to evaluate AI research in the UK|
|1988||New York-based AI hedge funds adopter D.E. Shaw is founded|
|2014||Man Group starts using machine learning for money management|
|2017||First AI equity ETF starts trading|
Businesses like Money Honey have sprung up offering digital-only advice services with online client portals and digital risk evaluation tools, allowing clients to choose their level of guidance or level of technological interaction.
Director Jane Hodges says: “If your business model is to build funds under management and manage those funds for an ongoing fee, then inevitably this will be overtaken by AI models. It is so much better to embrace this and find where the real value points are in the provision of financial planning and start to move to a model with more longevity.”
Artificial decision making
Returns on a standard £5,000 investment across different offerings by the seven major robo-advisers representing 90 per cent of the UK market vary significantly. In 2017, analysis from Boring Money comparing Evestor, Moneyfarm, Netwealth, Nutmeg, Scalable Capital and Wealthify, found that cautious portfolios returned the best results.
Wealth Wizards chief technical officer Peet Denny says robos can produce good results, but the challenge for product providers is building products that deal with complex-portfolio clients in the drawdown phase.
He says: “As clients get older and they accumulate more assets, their needs become more complex and it becomes much harder to fully digitise that kind of advice. Building products is a very manual process for those companies that do, and it takes a very long time.”
Denny says new AI offerings coming into the market will be able to double check where anomalies may lie. “AI can learn advice philosophy and policies and then tell if someone makes decisions outside that, as well as raise it with an adviser.”
Adviser view: Jeannie Boyle, financial planner, EQ Investors
We use an algorithm-based system that assesses risk and suitability, but always overlay that with a telephone call with an adviser, because that gives both the adviser and the client the peace of mind that it is absolutely the right thing to do. It is quite easy for it to spit out answers that aren’t quite correct or for outcomes to be slightly manipulated with AI-based systems. The combination of the system and speaking with an adviser allows clients to really understand what they are investing into and all the risks.
A 2017 Deloitte report on the future of automated financial advice in the UK concluded that: “Although the regulatory and legal requirements under which advice is given should be coded into the algorithm, rather than learned over time, machines will be able to have progressively more complex conversations with clients to, for example, understand their preferences or test their level of comprehension.”
In order to build intelligence within technology offerings, Deloitte says subject matter experts – human advisers – are still needed to assist with responses to unexpected situations until machine programming allows for nuances to be identified and catered for.
Adviser view: Lee Robertson, chief executive, Investment Quorum
AI is being viewed as a potentially revolutionary way that financial advice and the decisions made by investors within their financial planning can be improved. Published evidence compiled by UBS Evidence Lab appears to confirm this widespread desire to exploit the potential for better client investment decisions and outcomes. Of particular interest is using AI solutions for difficult areas such as risk assessment and raising the potential for success by removing human intervention and bias in risk tolerances and assessments. If this thorny area can be delivered or improved by AI many of us will cheer.
Who guards the data?
The security of AI algorithms’ pattern spotting and identification of client information continues to be a point of concern for financial planners.
Open Data Institute policy adviser Jack Hardinges suggests that better security will allow advisers and planners to better compare options across the market.
He says: “The right to data portability is likely to support individuals in switching providers of the same, or similar products and services. You can also port data to third parties.”
Altus Consulting says: “More work is needed in this area to enable safe and effective access to personal data, including anonymisation and de-identification, and making data more accessible in a way that protects privacy while creating a fair market.”
Denny argues that hybrid models where advisers work alongside AI solutions will provide the best of both worlds. He says: “This automated advice powers up a human adviser with products that allow the adviser to give advice eight to 10 times more quickly. Older people still tend to want to speak to a human, which is why we use a hybrid model.”
Involvement with regulators may also help advisers make clients feel more comfortable trusting AI-enabled offerings. Denny says: “The FCA needs to take, as its main metric, how well a product is serving clients, but parliament saying that they need to be involved to ensure customer outcomes are at least as good with automated advice is very promising. The FCA getting involved can only improve outcomes for customers and if robo-advisers are open about the fact that they are working closely with the FCA, then that gives customers and even other advisers confidence that they can trust an AI.”
Finance & Technology Research Centre director Ian McKenna says the challenge for the future is to embrace AI rather than compete with it. He says: “We need to learn from other industries where they are embracing how AI can help improve what
we deliver overall to our clients.”
Expert View: Simon Bussy, Domain Director – Wealth, Altus Consulting
AI will impact advice just like all other professions
“No computer will ever beat me.” So said Garry Kasparov, World Chess Champion, back in 1985. But it did, just 12 years later. In the past couple of years, Google’s brilliant Deepmind team has developed AlphaGo to teach itself – in just hours – to play the incredibly complex game of Go, and beat the best players in the world. Now imagine a similar AI initiative outsmarting the best tax experts, accountants or financial advisers and planners.
In broader financial services, we are already witnessing an increasing number of examples of demonstrable AI activity and success. Machine learning in the investment space is being used to deliver better returns, to replace portfolio managers, or to even focus on the financial advice process itself.
Although uptake and development of AI solutions in the UK to date has been tepid at best, there are increasingly creative applications of machine learning on more complex areas of the financial planning process and on the provision of fiduciary advice to create bespoke financial plans.
In time there will be no part of our lives that won’t be directly impacted by AI – specifically, machine learning – but we should remember that, at least for now, the best chess player in the world is not a human, or a computer; it’s a human and computer playing together.
Simon Bussy is domain director for wealth at Altus Consulting. Altus are just one of a host of expert speakers appearing at the Money Marketing Interactive conference next month. To book your place, click here.