The AI and robo-advice questions you should be asking

Automated advice, robo-adviser, machine learning, AI-enabled expert advice, robo-advice digital journey, hybrid machine/adviser solution … in a very short space of time the advisory sector has coined a significant amount of terminology for robo-advice.

And this is where the misunderstanding begins. Overused terms, concepts and clichés have started to cloud the issue and are trickling down to the end-user, potentially creating problems for firms and clients. While, for example, many financial service organisations may claim to use artificial intelligence (AI) to improve their processes and outcomes for customers, this could be an exaggeration and what they are in fact using is a narrow digital robo-advice journey.

Robo-advice has been under construction for a number of years, but it is only just now reaching its inflection point. As such, it is still a new concept for most people and many have had no direct experience apart from media speculation, or profiles of the few larger players that have reduced overheads and costs by making existing processes cheaper to run through automation.

The smaller players who are being far more creative and innovative in developing truly ground-breaking solutions driven by genuine AI do not have the marketing budgets to spend on mass communication. As the public gets to hear about these new innovations, they will start to realise the art of the possible when it comes to truly intuitive robo-advice.

Take the example of electric cars; five years ago Tesla was laughed at for suggesting that they could develop a mass-market electric car. Now all major manufacturers are spending billions on developing just that.

For financial advisory businesses looking to take the next step and integrate the latest innovations into their operations, there are a number of issues to be aware of both now and going forward. Being properly informed about these issues and asking the right questions of a provider can make an enormous difference in levels of investment.

What's the difference?

A large majority of robo-advisers are using decision-trees or simple chatbots to support their digital robo-advice journeys. This is not AI-driven advice but algorithms and models that conduct the initial investor screening, implement the investment strategy and monitor and evaluate the strategy where client portfolios are created using algorithms based on mean-variance optimisation.

The client screening process uses responses to online questionnaires - designed to identify clients' financial goals, investment horizons and risk tolerance - that are set up in the form of a decision tree. While these simple decision trees work well for simple cases, however, they can get very big very quickly as the advice becomes more complicated.

Experience has taught us this approach makes the advice difficult to analyse, difficult to build and difficult to test, leaving you with a system that takes a long time to build and is very resistant to change. Making updates to policies to cater for, say, regulatory or market changes are hard because changing one part of the decision tree will ripple through to other parts of the tree in ways it is difficult to control.

This is where AI can be a game-changer in the robo-advice journey. Being able to configure a robo-adviser to match the advice policy of a specific practice is significant and is achieved by using training data from historic or model compliant cases to develop the intuitive software required.

The AI system then continues to learn the house style by identifying patterns and rapidly configuring the advice software with the overall aim of reducing the time per case by weeks. Being able to learn which factors a particular advice firm emphasises when it gives advice is extremely powerful as each brand has its own way of delivering advice.

As an example, Turo, Wealth Wizards' AI system, effectively acts as a paraplanner, gathering data, scouring the market and presenting its findings to the adviser to be signed off. The paraplanner does most of the heavy lifting involved in giving financial advice but there is still a human being talking to the client at the beginning and at the end.

Turo currently works best when handling binary decisions, such as whether a client should transfer their defined benefit pension, take equity release or go into drawdown - although it will also flag up cases that fall into a grey area and need to be looked at by a financial adviser. As Turo learns more, however, cases such as these become fewer and fewer. Although some AI approaches like Deep Learning will do this in a way that is opaque - much like a black box - Turo uses explainable AI to create a model that can be easily understood by advisers.

Question to ask: During your discussions, ask if the suppliers have true AI capabilities or are they limited to providing you with a simple decision tree solution?

The first steps

And it will not stop there. Continual developments and enhancements in cognitive computing, big data and behavioural analytics will augment AI further and allow machines to sense, comprehend, act and learn.

This will quickly pave the way for technology to gather and understand complex client needs; propose and implement solutions based on historic client behaviour; and explain complex concepts in the context of each particular client. This could in turn pave the way for full automation of advice in the future and is significant for any firm investigating robo-advice options now.

There has to be the capability built into the current system to allow advisory firms to stay one step ahead of the sector, whether that is enabling a wider range of services - such as advice on investments, protection or mortgages - or attracting a larger number of clients and scaling costs so advice is accessible and affordable to everyone in order to build sustainable margins in the future.

Question to ask: During discussions, ask if your potential supplier has the vision and capability to expand your white-label robo-advice service?

The AI / human hybrid

Possibly the advice sector's greatest fear - and probably its largest cliché is that robo-advice will replace financial advisers completely. We just do not see this happening - in fact we believe the human-hybrid advice model is a powerful capability for financial advice firms.

While there are some in the sector who feel ‘fully-digital' is the only way advice can be made available, this may stem from the fact digitising is often seen as a way to cut costs. In many professions, however, digital is being used to improve the quality of the service being rendered, which is a big difference in emphasis.

We have found the human/AI hybrid gives the best possible customer experience, blending the efficiency, consistency and accuracy of an AI with the warmth, personalisation and care of a human.

Robo-advice is helping to fill a gap where there are far too few advisers - and it is our dream that, if we get this right, robo will help us train a new generation of advisers in the way chess computers have made a world-class chess education available to all. Human advisers will be with us for the foreseeable future. Only the mix of tasks they perform will change. Probably for the better.

Question to ask: A key area to explore is ascertaining how a potential robo-advice supplier can enhance your human adviser experience?

Inevitable rise in demand

As the general public learn more about robo-advice, they will be in the driving seat demanding the best-quality service and being able to differentiate between a firm that adopts robo-advice to cut costs from a system designed to provide excellence in value and service.

They will also be asking how is my data safe, how is it handled and what guarantees can you provide me with? Furthermore, they will be demanding an excellent human-computer interaction with a fully holistic advice experience - not just a PDF experience.

This is where a symbiotic robo-advice solution can enhance a firm's reputation through a seamless and painless human-client interaction. There is no doubt the role of the human adviser will continue to change as AI advancement will enhance robo-advice capabilities.

We anticipate advisers will spend more time on value-add activities in relation to their clients and increasingly less time on portfolio management as the roboadvice solution, driven by AI capabilities, does the heavy lifting. As this Accenture blog puts it: "AI will become the investment managers while advisers will become the relationship managers.

"AI capabilities will arm advisers with a comprehensive 360-degree view of the client through advanced technology that will empower them to proactively engage clients on what matters to them the most and on their terms. This satisfies the three main pillars of client intimacy: empathy, responsiveness, and self-sufficiency on the client's terms. Ultimately, enhanced customer experience will be the main differentiator for financial advisory firms."

Yet again, AI will make this deeper, more personal connection possible. As an example, Wealth Wizards is currently exploring new features for Turo, meaning enhancements such as voice recognition could be introduced in the near future. The possibilities are endless, with AI eventually supporting facial, in addition to voice recognition, in order to ascertain risk profiles based on behaviours.

Question to ask: Is your supplier able to talk about future enhancements and how they will achieve this?

This is a fast-paced sector and securing a roboadvice solution is a huge investment for any financial advice business. Unfortunately, the confusion over AI and legacy clichés have not helped companies seek the best longer-term solutions for their business.

Seeking out the latest in state-of-the art innovation and technological expertise is the answer and, while temporary disruption to your brand and business model is inevitable, adopting a white-label provider that will keep you one step ahead of the game can only deliver significant long-lasting benefits and upside.

This article originally appeared in Professional Adviser