Using the power of AI to help plan participants

May 13, 2020

A/AText size:AAA
Andrea Needham

Andrea Needham

Shannon Nutter

Shannon Nutter

Note: This is the first in a series of three articles detailing how Vanguard is personalizing artificial intelligence to drive better financial health, including factors that make AI worth the investment.

Artificial intelligence (AI) isn't coming to the retirement plan space. It's already here. Vanguard uses it every day in an effort to help participants own their future.

But what is AI? Most people immediately think of walking and talking robots. The kind we see in Hollywood blockbusters. The truth is simpler. AI is any system capable of certain behaviors associated with human intelligence, and these systems execute the tasks they were programmed to do. Sometimes we tell AI how to complete a task, such as when we want to automate a process. Other times AI is programmed to learn on its own to find more efficient solutions.

Shannon Nutter thinks regardless of what we're telling AI to do, the goal should be the same when it comes to retirement plans: to positively affect participants' financial health. Ms. Nutter, who currently heads the Vanguard superannuation proposition and customer experience in Australia, previously led participant strategy and development for Vanguard Institutional Investor Group and helped build the Vanguard AI offer and strategy.

"AI is powerful and its potential is immense," Ms. Nutter said. "But if AI isn't driving meaningful outcomes for clients, helping them improve their future retirement, then it's not worth the resources required to implement."

   We care about the participants we serve and want to get them back on
   track by using our AI techniques combined with behavioral science.

Helping clients when they need us most

While AI is helpful as a normal course of business, it can also help us better understand clients and how they react to unforeseen events, like the recent economic and lifestyle impacts resulting from the coronavirus pandemic.

"We are currently analyzing participant behavior and marrying that data up against public health data and government news feeds to identify potential trends in participant behavior," said Andrea Needham, department head of participant intelligence. "Our goals are to use this information to better engage with the participants and feed the data into future models."

Ms. Needham and her team leverage AI to look at thousands of participant actions, including voluntary suspensions (when a participant requests that their pre-tax deferrals get dropped to zero). By doing so, Vanguard can gain insights and identify trends that may inform suggested actions we can communicate when similar events occur in the future.

Voluntary Suspensions

Source: Vanguard data, as of April 10, 2020

"We're looking at current behaviors during these unprecedented times and hoping to gain insights so that post-pandemic we can appropriately reengage participants who may have had no choice but to suspend contributions," Ms. Needham said. "We care about the participants we serve and want to get them back on track by using our AI techniques combined with behavioral science."

Another area that has been impacted is participant-initiated pre-tax deferral rate increases. Vanguard often sees many participants increase the amount they save in the first few months of the year, to coincide with a raise. This year these increases are less frequent than normal, likely in response to COVID-19. "Participants may be sitting back and not making their 'normal' increases during this time of market volatility. We want to use AI to help identify trends with this population, so that once the markets stabilize, we are in a better position to motivate participants to save sufficiently for retirement," Ms. Needham said.

   … If our aim is to help as many participants as possible,
   then the law of averages fails you and them.

Moving beyond the law of averages

The market environment created by COVID-19 is one example of the law of averages not being enough to help as many participants as possible.

"We're an industry of averages," Ms. Nutter said. "Age, account balance, salary, contribution rate: Give us information like this and we'll make assumptions about what you should do to achieve a successful retirement. But if our aim is to help as many participants as possible, then the law of averages fails us and them."

Law of Averages and VG Participants

Using averages like demographics and plan data can create a gap. Participants could land in the same decile of a distribution but be in drastically different life situations. This is true in relatively calm times. Some individuals live more frugally than others. Someone could be caring for a loved one and taking on extra expenses. And it's especially true when you factor in the impact COVID-19 has had. Not only could participants be dealing with unexpected health care costs, they or their spouses could be facing reduced hours or the loss of a job.

These individuals will have much different behaviors than those less affected by COVID-19, and the law of averages can't tell that story.

Vanguard is using AI and behavioral finance to fill the gaps left by the law of averages. AI enables us to analyze more information about individuals than ever before to better understand their unique financial situations. By combining the output from our models with the techniques of behavioral finance, Vanguard can better motivate participants to take action and address their specific needs.

Up next

Of course, because AI is not walking, talking robots, there are humans behind every effective AI system. The next article in this series will focus on the importance of having the right people behind the scenes telling AI what to look for and where it can create efficiencies.


  • All investing is subject to risk, including the possible loss of the money you invest.
  • Diversification does not ensure a profit or protect against a loss.