Pension Funds Insider

Pension Funds Insider brings the latest pensions news and industry insights; from investment and governance updates to new mandate appointments and pensions regulatory information.

The prospects and challenges of AI in pensions administration

Image for The prospects and challenges of AI in pensions administration pension funds

Artificial Intelligence (AI) is rapidly becoming a focal point of interest among trustees, who are increasingly keen to understand its impact on pensions administration, its practical applications, and how they can leverage it to optimise service.

They are curious about how AI can transform their processes, improve efficiency, ensure greater accuracy, and enhance member satisfaction. While the potential benefits are significant, it's equally crucial to navigate this terrain wisely, acknowledging AI's limitations and applying it judiciously to avoid pitfalls experienced in other industries. By staying informed and strategic, we can utilise AI to its fullest potential, ensuring it serves as a powerful tool in advancing our industry and delivering outstanding experience and support to pension scheme members.

So, as we start to embrace this tech, let's have those important conversations about making the most of AI, ensuring we understand where it can benefit the industry, and most importantly pension scheme members, rather than a blind leap into the unknown.

Data management could be at the heart of how AI can transform pensions administration. We're surrounded by data in this industry, but the catch is that it's often dispersed, in various formats, or not even digitised. This is where AI steps in, offering a smart way to bring together and sift through these vast data pools. It's adept at identifying inconsistencies, streamlining data integration, and maintaining the integrity and uniformity of datasets.

However, the real challenge we face is dealing with data that's external or not readily integrated into our existing systems. Pension data is inherently complex and often requires extracting information from disparate sources, some of which may not be in electronic form. It necessitates converting legacy paperwork or unstructured data into a format that AI tools can understand and analyse.

Moreover, given the sector-specific nuances —where context can be everything— the demand is high for AI solutions that are not just powerful but also nuanced in their understanding. These solutions need to be capable of interpreting intricate details within the data, recognising patterns, and applying this knowledge responsibly.

Therefore, while AI holds immense promise for enhancing data management in pensions, its success hinges on our ability to tailor these technologies to address the unique demands and complexities of our field.

Calculations. While AI holds promise in automating pension calculations, enhancing accuracy, and efficiency, it's not yet capable of independently creating and writing complex calculation codes from scratch. The development of such codes still relies heavily on human expertise in analysing, interpreting, and understanding detailed scheme rules and complex data. However, once the specifications are established, AI can significantly aid in writing and refining code, identifying and correcting errors, and ensuring alignment with the calculation requirements. This points to a future where AI supports, rather than replaces, human expertise, offering tools to improve the reliability and efficiency of our processes. A key benefit of this for members and trustees is that through efficiency improvements the cost of calculation automation can be reduced, putting it in reach of more schemes across a greater spectrum of members.

Member experience. Enhancing the member experience is a key focus for pensions administrators, where AI, particularly through chatbots and automated response systems, is already making headway in providing timely support and information. Despite the efficiency gains, deploying AI comes with its quirks, such as the humorous incident where a chatbot recommended investing in time travel as a solution to increasing contributions. While these instances may bring a smile, they underscore a crucial concern: the implications of AI errors are far-reaching. Inaccurate advice from AI could misguide members in their long-term savings strategies or crucial decision-making, which could significantly impact their retirement planning. So, while AI has the potential to revolutionise member engagement with its rapid response capabilities, its integration must be carefully and cautiously deployed. The goal is to ensure that the guidance it provides is not only prompt but also precise and reliable, thereby upholding the integrity of members' financial futures and fostering their trust in the pension system. Right now, managing and controlling that balance is difficult and has limited returns, but watch this space as this is one area where rapid developments and scheme-specific learning will quickly evolve.

Rule analysis and benefit determination are among the most resource-intensive and expertise-reliant aspects of pension administration. AI presents a promising opportunity to revolutionise these areas by automating the complex processes involved in interpreting scheme rules and determining member benefits. While still emerging, AI's application in this domain can significantly expedite the analysis, reduce human error, and improve efficiency. However, for AI to be effective in this context, it must be adept at navigating the intricate web of legal and regulatory stipulations that underpin pension schemes – areas in which it still currently falls short, especially where language is vague, nuanced or contradictory. By enhancing AI's capabilities in understanding and applying these rules, pensions administration can achieve quicker, more accurate outcomes, ultimately benefiting both trustees and scheme members.

It's clear that AI is setting the stage for some exciting transformations in pensions administration. From the complexity of data management and the arduous detail of benefit calculations to enhancing member experiences and navigating the labyrinth of scheme rules, AI's potential is widespread. Yet, as we stand on this technological frontier, it's crucial to strike a judicious balance. We need AI solutions that are not just smart but also sensitive to the intricate details and regulations of pension schemes.

While AI's journey in our sector is filled with promise, it's still not, and won’t be for some time, a silver bullet. We must harness its power with care, ensuring it aligns with the specific needs and nuances of schemes and members. If we get this right, AI could be a game-changer, making pension schemes more efficient, reliable, and user-friendly. It's about enhancing, not replacing, human expertise, and paving the way for a future where pensions administration is not just quicker and cheaper but also more attuned to the individual needs of members.

So, as we venture forward, let's embrace AI with open eyes and open minds, ready to leverage its potential while navigating its pitfalls. After all, the goal is a bright one: to innovate pensions administration in ways that truly matter, benefiting trustees, administrators, and, most importantly, the members whose futures we're all working to secure. The journey of AI in pension administration is just beginning, and it promises to be an exciting ride.

Daniel Taylor, Client Director – Trafalgar House