Consumer Duty: what could it do to Pension Providers annual results?
Consumer Duty driven adjustments to profitability for life assurers come in two main areas:
1.????? Price and Value
2.????? Consumer Understanding
To date Price and Value has received more attention, not least because of SJP and the action they were required to take regarding their charges and approach to annual advice reviews. For product providers, many of which don’t provide advice, P&V is more about charges on old policies that are now deemed to be unfair and so have to be waived.
A key profitability measure is the IFRS AOP (Adjusted Operating Profit). This capitalises any change in assumptions over the lifetime of the existing block of business. Something that may have little cash impact in the current year can be magnified dramatically via the AOP.
Improving understanding is designed to help consumers make effective, timely and informed decisions. For example, it’s not enough to know that, if they withdraw money from their pension there will be tax to pay, the customer should be given an estimate of what is due and guided through other options if the potential tax bill makes them revise their decision.
What might this cost? ?MaPS, via Pension Wise, are the only guidance provider operating at scale. They delivered 243,000 appointments at a cost of ￡16.9m, or ￡70 each. This feels like the right order of magnitude for a process delivered by people. The equivalent cost using conversational AI is roughly 10-20% of this figure, call it ￡10.
Consumer Duty is new so the expected cost of complying will show up in the 2023 AOP. A strategy that envisages? speedy adoption of digital delivery will lead to a smaller adjustment than a plan to use people long term: For a company with a million pension customers the digital strategy might see a ￡10m adjustment in the AOP as opposed to ￡70m for human guidance.
Of course, having a strategy is not the same as delivering results. Our experience at Engage Smarter has been our early adopters have tried and failed to build a robust and accurate AI system themselves. For example, Chat GPT offers 65% accuracy for UK pensions questions, this can be improved to c80% with a quick project. To be a success, accuracy needs to reach the high 90s and that is where the difficulty lies, especially as early design decisions can be hard to reverse.
Turning back to the AOP – an ambitious strategy on digitising Consumer Duty will mean a smaller adjustment in the 2023 results. Conversely, a cautious strategy that relies on existing methods could lead to an adjustment of SJP like proportions and all that entails. Then again fail to deliver on the former and 2024 or 2025 results will look bad.