Skip to main content

You are here


What Artificial Intelligence Can Teach Us about 401(k) Plan Design

Data unearthed from tracking the most common questions employees have about 401(k) plans suggests that the complex nature of such accounts are a bigger barrier to saving than experts realize, according to a new white paper.

Vesting schedules, in-service withdrawals, emerging markets and lifecycle funds are terms commonly understood by 401(k) plan advisors, but these terms can seem like a foreign concept to the average plan participant, revealing a large disconnect.

Published by Dream Forward, “Retirement 2058 – What artificial intelligence reveals about 401(k) plans” shares insights learned from three years of developing the first AI chatbot dedicated to retirement, based on a database of employee conversations.

Authors Brendan Sweet and Grant Easterbrook, co-founders of Dream Forward, suggest that the structure and complexity of 401(k) accounts are just as big of a problem as the macroeconomic trends that affect America’s retirement system.

“One of the most interesting revelations from exploring our proprietary data set is the number of questions around understanding 401(k)s and financial terminology. Questions about tax rules, 401(k) basics and financial terminology account for about 63% of the questions our AI receives,” Sweet and Easterbrook write.

To better track pain points across the industry, they developed an internal metric called the Confusion Index, which follows the issues that create the most problems for employees. Not surprisingly, the topics representing the highest level of frustration and confusion include understanding investing concepts, rollovers, match/profit sharing terms and tax-related questions.

But the single topic that users struggle with the most was understanding the different withdrawal options, according to the data. Particularly challenging was comprehending the differences between in-service withdrawals, 401(k) loans, early withdrawal tax penalties, hardship withdrawals and in-kind distributions. “Participants really struggle to wrap their heads around the pros and cons of the multitude of withdrawal options,” the authors state.

Based on the observations, the firm offers several suggestions to drive better outcomes.

Rollovers: Because many of the AI’s questions pertain to help with rollovers, the U.S. should update the rules and regulations around rollovers, the firm contends. The authors emphasize that consumers need help, especially with the paperwork needed from the prior provider that holds the account they’d like to roll over. “We see this pain point across all ages and income levels. Other countries (like Australia) have passed legislation to make the rollover process much more streamlined,” Sweet and Easterbrook write.

Simplification: The U.S. should drastically simplify 401(k) rules and regulations or build tools to help overcome the barriers caused by these rules and regulations, the authors further suggest. They note that overly complex regulation and jargon intimidates employees, particularly as 401(k) plans have grown more complex over the past 40 years. A prime example is users struggling to understand the difference between hardship withdrawals, in-service withdrawals, 401(k) loans and early withdrawal tax penalties, which is especially problematic at the lower end of the income spectrum, the paper notes.

Technology: According to the paper, technology needs to proactively reach out and talk to users during critical decision points. The technology would be designed to support financial advisors, not replace them, by giving them the data and intelligence they need to do more targeted outreach to employees.

“Without the AI intervening, users were all too happy to speed through the enrollment process, ignore warnings and prompts, and make a suboptimal decision for their long-term finances,” Sweet and Easterbrook emphasize.