Author

Mark French

Mark French

Associate Director

Willis Towers Watson





AI may be on the cusp of driving our cars and diagnosing our illnesses, but it is also about to enter the DC pensions market in a big way. In this article, Mark French considers how behavioural science and machine learning are coming together to radically and powerfully change the way in which pension schemes can communicate and engage with their members.

The age of intelligent machines is here. Whilst they are not yet walking among us in humanoid form, they are quietly infiltrating everyday life. They are the algorithms and the machine learning that increasingly sit behind many of the digital services we interact with.

AI, no longer science fiction

At a basic level, AI is a computer programme that is able to adapt and respond to specific tasks through interaction and statistical models. AI progressively improves its performance by learning similarly to a human.

We are already experiencing AI being used to manage communications from mainstream online retailers. If you have bought something from a website such as Amazon before, you might notice that it will begin to show you other items that may be relevant to you. These automated messages are designed and managed by an algorithm that is looking at your personal data and similar data produced by millions of other customers. It is making an informed decision about what items you may want to buy next based on probability.

We are seeing DC providers investing heavily in AI so they can manage their communication strategies in a similar way to big online retailers
 

We are seeing DC providers investing heavily in AI so they can manage their communication strategies in a similar way to big online retailers. The effect this may have on member engagement could be profound.

We live in a data rich world

The term ‘big data’ defines the ability to use large volumes of information to reveal patterns relating to human behaviour and interactions.

We are revealing more and more about our behaviour when we are interacting with the digital world and this data is being captured. DC providers are aggregating customer data across multiple touch points and building up unprecedented insight into pension scheme member behaviour. However, this insight is only valuable if it can be used in the right way.

Personalisation

Segmentation is the ability to use data to identify groups with similar characteristics. Whilst not a new concept, segmentation has arguably been underutilised when tailoring benefit design and scheme communications for DC schemes.

Our 2017 study of UK workers, The savings psyche of the UK, concluded that there are 6 separate personas into which UK employees can be categorised when looking at their attitude to saving. By understanding the wants and needs of individual groups, segmentation enables personalisation that will embrace these different behavioural traits and personas.

Behavioural science shows that individuals are far more likely to respond to personalisation and this can be further amplified with behavioural communication techniques such as ‘nudge’ and ‘peer group’ messaging using positive reinforcement as a way of influencing our behaviour. For example, HMRC revealed that issuing people in arrears on their taxes reminders that contained phrases such as “9 out of 10 people in your area are up-to-date with tax payments” increased the proportion of people who paid their income tax before the deadline. These techniques tap into our most basic traits of apathy and wanting to fit in.

Hyper-personalisation

Hyper-personalisation combines the sophistication of AI with the power of behavioural science to create a highly personalised and highly relevant communications capability
 

Hyper-personalisation combines the sophistication of AI with the power of behavioural science to create a highly personalised and highly relevant communications capability.

Where personalisation is able to tailor communications towards characteristics such as pot sizes, age and risk appetite, hyper-personalisation is about using AI to adapt to an individual’s decisions in real-time. In DC pensions, this means understanding an individual’s engagement with the employer’s scheme by recording and analysing multiple data points about them and their interactions in order to construct relevant content from the available communications collateral.  The AI is able to look at the individual’s behaviour in response to specific targeted messaging based on segmentation. It analyses whether it was read, whether links were followed and whether any action was taken. It learns from the results so that it can adjust the content for the next message which could act as a nudge or a call to action.

The AI sits behind a multi-channel communication strategy and can decode how an employee may want to engage with the scheme through the data that it has gathered. It makes a decision about how, and when, it is best to communicate to that individual to optimise engagement. Scheme sponsors may wish to leverage this advanced capability to meet specific objectives, for example they may want members to review their contributions or investment choices, or perhaps consolidate pots from an old scheme into a new scheme.

Bots

AI is not limited to managing campaigns. It is also powering online assistance and guidance services through webchat. So called ‘Chatbots’ are increasingly part of a multi-channel digital strategy. They are the virtual assistants that, through online or voice interactions, can navigate choices and accurately select the appropriate response suited to the customer. You may have experienced these when calling your bank or interacting with an online webchat service and may not even have noticed at first that you were interacting with a machine rather than a person. In future, machines such as these could be equipped with conversational skills and be able to express empathy. They might just be able to pass the ‘Turing test’.

Developed by Alan Turing in the 1950s, the ‘Turing test’ is a test of a machine's ability to exhibit intelligent behaviour equivalent to, or indistinguishable from, that of a human.

The future is bright

The age of intelligent machines is here. Combining proven behavioural science techniques with AI powered communications looks set to genuinely start to break down apathetic behaviour among DC members.

Artificially intelligent communications could be one of the most powerful tools to hit the DC pensions landscape in decades
 
This technology may be the key to solving some of the biggest challenges facing DC schemes, such as delays in engagement leading to poor retirement outcomes. Artificially intelligent communications could be one of the most powerful tools to hit the DC pensions landscape in decades.