Innovation: Is AI in the pension industry at an inflection point?
‘A look ahead to 2023’ download our full report below
Artificial intelligence (AI) and machine learning tools capable of analysing large data sets are becoming increasingly understood and accessible. With their help, better solutions can be developed, and processes automated that previously required a lot of human ‘handle turning’.
Given the swathes of data the
Felix Mantz, Director
pensions industry holds, we hope that
2023 brings more AI-led innovation.
Looking at covenant within the defined benefit (DB) space first, we expect to see more AI and related analytics tools help improve outcomes and drive efficiencies particularly for large multi-employer schemes and small schemes.
As examples, for multi-employer schemes with 10+ employers, AI can help by undertaking proportionate reviews of the participating employers and summarising the findings on an interrogatable platform. This is particularly helpful for industry schemes or nonassociated multi-employer schemes, including Local Government Pension Schemes (LGPS) or DB master trusts, which often have hundreds of employers.
For small schemes with assets under £50m (i.e. broadly half of the c.5,500 UK DB schemes), AI can help by providing proportionate covenant assessments in a more cost efficient manner. This allows small schemes to benefit from external benchmarking and validation that robust covenant advice brings, but without the costs of in-depth covenant analysis by a team of dedicated professionals.
But the applications of AI in pensions are far wider than just the covenant sphere. For example, defined contribution providers may want to analyse member decisions to improve the design of their product and platform. Administrators may want to use AI to help identify and fix data issues, in particular in the run up to a DB scheme bulk-purchase annuity (BPA) transaction.
Asset managers and investment consultants may also increasingly turn to AI to analyse unstructured ESG information to improve reporting and portfolio allocation.
That being said, two challenges come to mind for the adoption of AI based tools, both for pensions and in general.
The first is a practical consideration – much like for the aforementioned BPA transactions, good data is key. 80% of the work to develop an AI solution usually revolves around data collection and cleaning, rather than the ‘sexy’ machine learning applications.
The second challenge is a more emotive apprehension of AI, often rooted in uncertainty over how good (or not) it is. Within a narrow and well defined domain, supported by lots of data, AI tools can be incredibly powerful – but their output needs to be interpreted and their limitations
understood. They can help automate the boring stuff and free up some of our time, but ultimately still require humans to point them in the right direction.
With this in mind, we believe the industry is at an inflection point – with the use of AI set to increase. 2023 will be an exciting year – and we will do our part to contribute to the innovation that AI can bring to the pensions sector.
A look ahead to 2023
A year of reflection,
consolidation and
recasting strategies.
A look ahead to 2023
"*" indicates required fields
