Transforming retirement planning with big data and AI

6 Apr 2021

In the dim and distant past, when I was still an IFA, a couple of work colleagues did rather well switching their investment bond funds on a near-daily basis.

They had a pretty formulaic way of deciding which fund to be in on what day. I don't remember too much of the detail. If I say the trades were placed using fax machines then you can work out how many decades ago.

The process went something like: call your granny, ask her to look at Ceefax on the TV (if you're too young, look it up), get her to read out the overnight movements of some key indices from around the world, make a decision based on the data to switch funds, send off a fax to the life company and do the same on every working day.

They also made sure they were always in the property fund on the day the monthly accounting happened. It always amazed me how so few data points used in the right way allowed them to game the system. Small beneficial decisions, based on some simple rules which were ‘more right than wrong', made a material difference.

It reminds me a little of the glory years of British Cycling following the 2003 appointment of David Brailsford, now Sir David John Brailsford CBE, as performance director. Famously Brailsford said: "The whole principle came from the idea that if you broke down everything you could think of that goes into riding a bike, and then improve it by 1%, you will get a significant increase when you put them all together."

The list of adjustments is too long to list here but included hiring a surgeon to teach each rider how to wash their hands to reduce their chances of catching a cold, establishing the right mattress and pillow combination to ensure the best night's sleep for each rider, and even painting the inside of the team vehicle to make dust easier to spot. Small beneficial decisions, based on some simple rules which were ‘more right than wrong', made a material difference.

Big Data

Roll time forward and we are in the era of big data, and hugely powerful computing technology that crunches that data faster than humanly possible and can act on it instantly. In the sporting arena, individuals spend the majority of their training and competitive time connected to Global Positioning Satellites - measuring distance travelled, acceleration, explosive power and who knows what else. Every calorie consumed is accounted for by food type, nutritional value, energy release and recovery requirements. Small beneficial decisions, based on some simple rules which are ‘more right than wrong', making a material difference.

So how are we doing in financial services and in retirement planning and provision in particular? What are those small beneficial decisions, based on simple rules which are ‘more right than wrong', that can make a material difference to someone's retirement?

Many artificial intelligence (AI) and machine learning (ML) applications will remain largely invisible to the end customer. Yet incorporating AI technologies into complex financial forecasting and planning have the potential to be transformational, because the long term nature of retirement planning affords the time for a series of tiny improvements to accumulate great wealth.

Change The World

The ability of AI to improve predictive (what will happen) and prescriptive (the best course of action) financial forecasting processes will change the world of finance management. Currently, many financial forecasting and planning processes are manually intensive and suffer from inherent human biases, as predictive models may be ‘tweaked' to generate favourable (or expected) outcomes.

AI understandably underpins significant volumes of fund management decisions. That same rationale will no doubt increasingly extend to individual financial planning, influencing decisions at the personal account level. Shaping an individual's income based on analysis of historic spending patterns resulting in pension assets remaining untouched for longer is just one basic application that, over the duration of a person's retirement, could generate those marginal gains.

During the ever longer accumulation phase, what difference could AI driven nudges that influence pushing micro-savings into retirement savings have? A lifetime's worth of rounding up retail purchase payments, with compound interest growth, adds up to a tidy sum.

Nudges can be driven by AI ‘noticing' when spending patterns change and there may be a temporary increase in disposable cash. The in-retirement equivalent of this kind of nudge may be even more valuable over the long term - not taking income when it isn't needed can be hugely beneficial particularly during a market downturn. Mitigation of sequencing risk (the evil twin of pound cost averaging) is not to be underestimated.

Every part of the chain in pensions is under pressure to justify and deliver value for money. Being able to demonstrate how you are harnessing new technology to benefit clients is a great way to show that value for money is not just about having lower fees than the competition.

Digital personal assistants have come a long way since the Microsoft Office paperclip. Expect increasing use of chatbots in retirement savings propositions that improve digital customer engagement in an effective, low-cost way.

The next generation of chatbots are learning the type of language and phrasing that best resonates with an individual customer. The volume of data supporting the increasingly complex ML driven algorithms will be a game changer in this space - providing timely messaging and tactical responses to the smallest micro and macro changes.

ML algorithms are learning and adapting with each new data point and they become more sophisticated and accurate over time - so things can only get better! Embracing AI and ML as key components in the on-going advice process need not be a threat to jobs - more like a tool to allow humans to focus on the parts of the customer engagement that humans do best, in the comforting knowledge that Big Data has their back.

Those that successfully harness AI and ML as business tools will do well over the long term. Rather like the cycling world, others that ignore the potential of small iterative improvements from rules that are ‘more right than wrong' may be surprised when they are left behind by their competitors.

by Andrew Martin, Chief Distribution Officer at Dunstan Thomas
Click here for the full article on Professional Adviser.
Follow Andrew Martin on Twitter: @ajmdunstan or read Andrews's previous article here.

Andrew Martin
Chief Distribution Officer at Dunstan Thomas