Overview
Artificial intelligence (AI) now pervade our daily lives, from personalized online shopping algorithms to crowd-sourced navigation assistance. It is all around us, making our daily customer experiences more personalized and our lives a bit less frustrating. Its proven value is now being applied to employer-based retirement savings programs on a global scale. AI is increasingly influencing the structure and delivery of retirement savings plans by providing plan sponsors with better tools and information for monitoring portfolio managers and giving plan participants a richer and more customized user experience.
Financial decision-making in retirement savings plans
For most people, saving for retirement is the most important sequence of financial decision-making processes and actions in their lifetime. The rise of personal retirement savings plans and the associated shift in responsibility to the individuals have been accompanied by advances to help those individuals make better financial decisions and mitigate the associated anxiety.
The challenge and opportunity facing plan sponsors today is how to use advances in AI technology to augment the foundation provided by participants and behavioral finance so that retirement savings plans deliver even higher performance for those entrusted with oversight responsibilities.
AI and retirement savings plans
AI can mean different things to different people but for present purposes we will focus on mining data, identifying patterns in behavior based on that data and developing responses that can improve the user experience and decision-making abilities for retirement plan sponsors and participants. One conventional example would be the “chatbot”, which enables plan participants to hold an on-line conversation with a support “person”— in reality an AI-aided computer.
Behind the scenes an important distinction is taking place. The source and relevance of the data underlying the AI-based communication can vary from broad based to highly personalized. In the context of a retirement savings plan, the sourced data can be categorized across three layers of personalization and effectiveness:
Depth and complexity of data sources
Layer 1 : Highly generic sourced data and information to answer questions transcending the plan. For example, answering general questions about tax regulations related to withdrawing retirement assets.
Layer 2 : Plan-specific information forms the basis for AI interactions, whether relating to plan design or overall participant profiles. For example, information about how plan participants can take a loan using retirement funds Or, for the plan sponsor, AI-based assessment tools to monitor portfolio managers’ adherence to their investment mandates.
Layer 3: The most advanced application of AI would utilize individual plan participant information including investments, plan participant age, savings rate and retirement asset goals, to prompt highly personalized interactions and decisions such as presenting optimal withdrawal rate information based on the participant’s specific circumstances. AI can bring value to not only the participant engagement process but also three other key building blocks of a well-designed, delivered and monitored retirement savings plan: plan design, participant engagement, plan governance and investment strategy.
AI enabled building blocks
Below four building blocks helps how AI capabilities can be used to enhance each of them.
Four retirement savings plan building blocks
Plan design
- The retirement savings plan structure perfectly meets all plan objectives, delivering 100% participation in the appropriate risk-adjusted investments at savings levels that will maintain an employee’s standard of living through retirement.
- Even taking into account the insights offered by behavioral economics, plan sponsors can often fall short of this goal. Capturing and analyzing information underpinning plan design weaknesses at both a generic and a personalized level can translate into improved results.
- For example, examining the profiles and actions associated with participants taking advantage of the corporate savings match can lead to design adjustments that better serve participants’ long-term interests.
Participant engagement
- We know that without participant engagement, the best behaviors and results for the plan participant are hard to achieve. Plan communications and delivery methods must resonate with participants; otherwise, emails or texts will be quickly caused to the trash bin icon.
- The most common AI-related participant engagement tool is the chatbot. Chatbots are a means for consumers to interact, ask questions and get personalized feedback from a computer based on available data. Amazon’s Alexa is one well-known example. Offering positive reinforcement for actions that strengthen financial security and persuasive communications when a poorer decision is being considered are all ways AI can enhance plan participant engagement, encourage informed investment decisions and, ultimately, improve retirement outcomes. AI can enable the chatbot to evolve from a reactive service to being a proactive device informed and activated by broader participant milestones such as salary raises.
- Imagine, for example, a participant trying to decide between two different investment strategies. A chatbot or educational primer explaining the trade-offs and risks of the two options based on data collected on the participant’s personal goals could be highly influential towards a positive outcome.
Plan governance
- A core component of strong plan governance is the plan sponsor’s ability to fully execute their fiduciary obligations. This requires aligning plan guidelines and objectives as closely as possible with the employee population, refining over time based on actual data collection and analysis.
- AI technologies empower plan fiduciaries to take more proactive steps to fulfill their obligations. For example, participants experiencing a positive relationship with their retirement savings plan may be more likely to remain with the plan after leaving the company or retiring.
- Plan sponsors can also use AI to better align products and services to participant needs by developing more sophisticated plan participant profiles.
Investment strategy
- AI technology can aid sponsors with the on-going design and revision of investment offerings to best reflect each participant’s individual profile. For example, a plan sponsor can analyze participant withdrawal data to create target-date offerings that are best suited to the particular needs of their plan’s participants.
- AI has also enabled more sophisticated valuation methodologies for alternative asset classes such as real estate, making it feasible to include them in a daily-traded retirement plan environment and expand diversification options for participants.
5 steps to an AI-enabled Retirement Savings Plan
The following five steps provide a good roadmap for plan sponsors evaluating how to employ AI to improve the design and functionality of a retirement savings plan.
Step 1 :
Plan objectives related to plan participant behavioral characteristics most conducive to AI applications could include those related to increasing:
• Participant engagement and tool interaction
• Savings contribution rates
• Confidence level in long-term financial security
• Financial awareness and literacy
Step 2 :
Analyze plan participant action or inaction that may result in suboptimal results.
Step 3:
Types of information and knowledge might improve key metrics and outcomes : Data analysis and participant surveys aligned against gaps in performance such as under-saving and poor asset allocation, combined with insights from a “confusion index” can highlight priority areas.
Step 4:
Participants have preferences for where and how they want to interact with the retirement savings plan. Using technology to address those biases can help improve engagement levels.
Step 5:
Plan sponsors can use AI to conduct sophisticated sentiment analysis and then customize plan communications accordingly to maximize positive participant responses.
