How artificial intelligence could lead to a customer service revolution in pension savings

How do you rate customer service in pensions in the UK?

If you rate customer service experiences from great to awful, how high up the list would you place dealing with pension schemes? For many people, pensions are difficult to understand and time-consuming to deal with. Paper forms, large amounts of information and a whole jargon of technical terms.

Does it have to be like this?

Another world is possible

Imagine a future in which:

• dealing with your pension scheme is as simple and satisfying as shopping with the best online retailers;
• information is targeted and automatically provided at key points in your life;
• processes are simplified and sped up, so that you get what you need as soon as possible.

Let me paint a picture of what this world could be like. This pension scheme of the future has both human and digital employees. Human employees take up front office positions, dealing directly with scheme members as pension advisers, nurses and employer contacts.

The scheme also has a busy back office function, dealing with member queries, sending out targeted communications, processing member requests, paying out critical illness and death benefits, making actuarial calculations and updating records. The difference between this scheme and many others is that this back office is run entirely by digital employees.

How artificial intelligence powers a customer service revolution

Whether you call them bots, robots or digital employees, they are embedded at the heart of the scheme’s activities.

Imagine that the scheme pays out a lump sum if a member is diagnosed with a critical illness. In traditional pension schemes, this could involve the member having to fill in pages of forms, provide access to their medical reports and submit to medical assessments by the scheme’s own doctors.

In our pension scheme of the future, the member simply goes online or uses the scheme’s app on their mobile phone. On the critical illness page there are just three options:

• a drop down list of illnesses;
• a box to type in the hospital where the diagnosis was made (which, of course, intelligently guesses the name of the hospital after you’ve entered a few letters); and
• a calendar for you to provide the month and year of the diagnosis.

You can then submit your online health records or provide a digital power of attorney to the scheme. In the latter case, the scheme’s bot will use a secure communication channel to write to the hospital asking them for the relevant medical records. Hopefully, the hospital will respond as quickly as the scheme, but, if it hasn’t written back within four days, the scheme’s bot will nudge the hospital and tell the member that the hospital is being slow.

In the meantime, the bot will check the exact entitlement for the member and prepare a payment to be made, so that no time is lost when the medical records are released.

Eventually, the hospital sends the medical records to the scheme. The scheme’s bot checks that the details on these records matches what the member has said. If it has any questions, it can ask one of its human colleagues to intervene – not even the cleverest bot can read all examples of handwriting!

If everything matches up, the scheme’s bot will make a payment and send a message to the member confirming that this has been done. It will communicate in the member’s preferred channel, sending an email, a letter or a message on the scheme’s app depending on how the member communicates with the scheme.

The bot’s work is almost done. The last things it needs to do are to:

• arrange for membership with a relevant health charity (e.g. the British Heart Foundation for a heart condition); and
• send a message to one of the scheme’s (human) nurses to contact the member.

Because all of the back-office administration has been carried out by the scheme’s bot, the nurse now has time to provide the human touch of empathy, advice and care. Rather than reducing the number of human employees, the introduction of bots has freed them up to spend more time on things that make a real difference to members.

Bot of all trades

The scheme’s bots are not just employed on critical illness claims. They are able to scour public records to initiate communications and payments. For example, in this case, the national births, marriages and deaths registry is open, digital and searchable. The scheme’s bot goes in every day to see if there is a change in status for any of its members.

A member’s death will initiate the death benefits process. The system is programmed to be sensitive. It could initiate payment within a matter of hours but, instead, it holds off sending any communications until at least a week has passed. But once it is on the case, the bot can:

• check all of the details against the national registry;
• determine if there is a spouse;
• review expression of wish forms; and
• deal with the relevant probate agency.

Data is at the heart of what this scheme is able to do. The scheme is able to gather a wide range of data about the member from both internal sources (e.g. pension contributions, payroll information, employer details, length of service, benefit payments, what the member has looked at online or on the app, communications sent by the member and sent to the member etc.) and external sources (e.g. national databases of house purchases, central registers of deaths, births and marriages and credit scores).

With all of this information, the scheme does not need to send blanket mailouts or generic member communications. Instead, it sends automatic, targeted and personalised communications that are relevant to the member at specific times. For example, if the system becomes aware that the member has had a child, it will prompt them to increase their private health insurance. If the member has bought a house, it could suggest suitable insurance products. If the member receives a bonus or a pay increase, the system might nudge them to save some of that extra income into their pension.

A real-life example

What is truly incredible about all of this is that this is not a pension scheme of the future. These are all real-life examples from PensionDanmark, one of the 50 largest pension funds in Europe.

PensionDanmark is a member-owned, non-profit, labour market pension fund with €36 billion of assets under management. It has 713,000 members employed in 26,400 private and public companies. And it has been carrying out these automatic processes for the past three years.

The results have been extremely high level of member satisfaction. Between 93% to 96% of members report being satisfied or very satisfied. In addition, it is able to provide better service at lower costs. Cost per member has fallen from €55 per year to €40 per year over the past ten years.

Will we see anything like this in the UK?

Yes and no.

Pension dashboards will be the catalyst for schemes to produce data that is high quality, accurate, standardised and digitised. Once that is in place, the options available to members should increase. It should also be the fuel that stimulates a growing industry around pensions technology.

Not everything that is possible in Denmark will be possible in the UK. The UK does not have the same central registries or health record systems. But we can move closer to our Nordic neighbours in terms of customer satisfaction and efficiency.