Artificial intelligence (AI) seems to be everywhere these days. After ChatGPT became a household name in what seems like overnight, it would be difficult for industries to overlook this technology – and the pension industry is no exception.

As much hype as there is surrounding AI, it’s not as new as some are claiming it to be. We already use AI from morning until night – and we have been for years. Did you Google something today? If so, you used AI. Google is arguably the most widely used AI in the world (Malone, 2022). Watch something on Netflix? Any movie recommendations Netflix gives you are generated by machine learning technology (a core branch of AI) that learns your preferences as you use the streaming service.

In fact, the concept of machines learning like humans has been around since the 1950s. By the 1970s, programs could learn from examples and answer questions – much like Siri or Alexa do today (Winston, 2022).

Many of the AI applications we’re already familiar with can also be used to enhance communication with pension plan members. Plus, advances in generative AI can help pension communications teams be more efficient when it comes to creating and designing engaging content. 

Let’s unpack four of the most significant uses of AI to help pension communications teams thrive.

1. Content creation – Creating meaningful content is becoming increasingly time consuming. Members often require specific information depending on factors such as how near (or far) they are from retiring. This results in a need to produce more targeted content. Generative AI refers to systems that can produce text, images, audio, video and other content based on patterns learned from thousands of data points. 

Generative AI tools like chatbots fall under the natural language processing branch of AI – and this branch has made impressive advancements in its ability to create more realistic, accurate and human-like outputs. As a result, it is the branch of AI that we’re hearing about almost everywhere these days.

Generative AI can be a game changer in terms of the time it saves you in drafting compelling content to get your members excited about having a high-value benefit like a pension. However, there are some best practices to keep in mind when using generative tools for content creation:

• Revise and refine – Generative AI tools offer a great starting point to help you go from a blank page to a draft in seconds – but let’s emphasize the word draft. AI should help you create your first draft, not your final draft. You still need the expertise of people to edit the outputs, check for accuracy and add an emotional perspective to the content. 
• Anonymize your data – Instead of using the names of actual companies or people in your prompts, use pseudonyms. The prompts and questions that you ask chatbots are often saved and used to train their large language models (LLMs) by default. This can introduce privacy issues. However, most chatbots have the option to opt-out of data sharing – a good idea if you plan to share sensitive information.
• Use disclosure statements – If you’ve used AI to create content like video or images, it’s important to disclose this to your members. Be transparent. This helps establish trust with your audience and your brand.

2. Summarizing information – Another key strength of AI is its ability to quickly summarize complex information. Communications teams can use generative AI tools like ChatGPT, Copilot and Claude to significantly simplify and shorten pension booklets and reports. 

Generative tools also excel at summarizing and finding trends in analytics and survey data – work that would take hundreds of hours to complete by humans. 

By speeding up your team’s workflow, generative AI frees up valuable time for you to focus on high-impact, strategic tasks that drive results.

3. Hyper-targeted communications – In the defined contribution (DC) pension space, tailoring information encourages individuals to take small steps to make better choices for themselves and their financial futures. 

Additionally, members now expect regular access to digital information that is targeted to their own retirement planning journey – much like the personalized digital experiences they have with tech giants such as Facebook or Spotify.

The machine learning branch of AI can significantly help pension communicators more precisely target and automate information. Most plans have pension administration systems that are jam-packed with member data. Deep learning algorithms can quickly analyze the data and group members according to similar features; in fact, this is one of the most significant ways AI is being used to target communications across a variety of industries. These segments can then be used for targeted campaigns, such as providing members grouped near retirement age with information about de-risking their investments.

4. Real-time recommendations – Machine learning is also gifted at making real-time recommendations. Many product recommendations you see on Amazon and other e-commerce platforms are driven by machine learning that detects patterns in your interactions with the online service. 

Based on your members’ behaviours and interactions with online channels like your website, email campaigns and social media, pension communicators can use machine learning to recommend relevant retirement planning services, articles, videos and other content that your members may find helpful. These recommendations will become more accurate over time as the system learns more about your members. 

Ethical concerns

As much promise as there is surrounding AI to help pension communications teams thrive, the technology is not without risks. It is important to consider the ethical concerns that could have a negative impact on your organization.

• Data security considerations – Data security is a serious ethical concern that could arise for pension plans. For instance, a machine learning algorithm would need to analyze thousands of data points in the membership database to find patterns. This raises data security concerns since the machine would have access to sensitive information that – if ever leaked into the wrong hands – could lead to malicious acts such as identity theft.
  
  It is important to make sure your organization has the proper data security infrastructure in place and that members are informed about how their information is collected and used by any AI systems.
  
  
• Privacy considerations - The protection of both your members’ and your organization’s privacy are a real concern when it comes to implementing an AI roadmap. This risk is high when using generative AI tools like chatbots. 
  
  If your plan provides a chatbot on its website to answer member questions, you have no control over what information members might input into the bot. They could provide SIN numbers, birth dates (their own or a beneficiary’s) and other personal identifiable information. How will your organization ensure this information is kept confidential? 
  
  The risk also stretches to your organization’s protection of information. Is your team using a chatbot and inputting confidential business information into the system? Many chatbots store and use the inputs you give them as data sets to further train their bots; this can create risks if a bot is asked to summarize a confidential strategy. Make sure you check the terms of use and privacy parameters of the chatbot you’re using so you know how it’s handling your data.
  
  
• Unanticipated behaviour - Unanticipated system behaviour is also a very real ethical risk, specifically for personalized content initiatives using the natural language processing branch of AI (i.e. chatbots). 
  
  If a machine provides content recommendations that are not appropriate for a member, the plan could be held liable. For instance, if a member of a DC pension plan on the cusp of retiring received content recommendations that encouraged aggressive investing behaviour and acted on it, the member could lose a significant amount of money if the markets were to suddenly drop. It’s always a good idea to keep a “human in the loop” to review any outputs provided by the AI system to ensure they are accurate and factual.

Conclusion

AI holds great promise for pension communications teams to enhance efficiencies when it comes to creating engaging content, summarizing complex information and data, and developing hyper-targeted and personalized campaigns that can help boost member retirement outcomes. There are, of course, many other uses of AI that can benefit your pension organization. When thinking about potential uses for AI in your organization, a good question to ask is: What tasks should computers do, and what tasks should people do? Humans are stronger than computers when it comes to interacting with other people and showing empathy – these traits are very difficult to program into a system. Computers, on the other hand, excel at remembering and making sense of huge amounts of data.

Together, humans and computers can improve the experiences we provide for our members. If your organization has the proper data security foundations in place to support this technology and your team is educated on responsible use of AI tools, it can be a powerful addition to your communications program for the benefit of members. 

References

Malone, Thomas (2022). Artificial Intelligence: Implications for Business Strategy. Interactive Video Set: Video 1 Part 1 Transcript. MIT Sloan & MIT CSAIL.

Winston, Patrick (2022). Artificial Intelligence: Implications for Business Strategy. Casebook Video 1 Part 1 Transcript. MIT Sloan & MIT CSAIL.

Nicole Quintal, Manager, Brand and Digital Engagement, Co-operative Superannuation Society (CSS) Pension Plan

Nicole Quintal is an award-winning communicator with over 18 years of experience in the journalism, marketing and communications fields, including 10 years in the pension industry.

Inspired by the tech and fintech industries, Nicole is passionate about exploring innovative ways to elevate communications with plan members. Nicole and the CSS team recently launched new initiatives, including new targeted video content, a new website, and personalized information delivery for members.

Nicole holds a Diploma in Journalism from MacEwan University, a Bachelor of Professional Arts in Communication Studies (Honours and Distinction) from Athabasca University, and a Pension Plan Administration Certificate (PPAC) from Humber College. She also recently completed programs on AI and blockchain technology through MIT’s Sloan School of Management.

In her spare time, Nicole loves hiking in Northern Saskatchewan with her German Shepherd, Yuri.