Finding the needle in the haystack
Over the coming year, we're expecting the rise in complex pension advice claims to continue, and more self-invested personal pension (SIPP) operators to fail. The more complex the claim, the longer it takes (and more it costs) for us to handle, but we're always looking for smarter and more efficient ways to work - getting any compensation to our customers as soon as possible.
We recently shared how FSCS has been embracing artificial intelligence to help process claims. Another innovation we've introduced recently is a new data lake and search tool called the ‘Resolution Data Lake’. It might sound technical, but the idea is simple - and effective.
What's a 'data lake'?
A data lake is a vast pool of raw data, that has no structure or defined purpose. Think of it a bit like that drawer we all have at home where we keep those 'important' bits of paperwork - but dread having to try and find anything in it!
Storing information in a data lake means everything is in one place, which is handy, but it's crucial to make sure it's easy to find what you need.
Why did we need a new search tool?
When a firm goes out of business, it's common for us to receive over 1,000 pages of documentation for each customer, particularly in relation to SIPP operators which are usually the most complex claims to deal with.
Often these documents are formatted as PDFs or images, or even as audio files in the case of phone recordings. This makes searching through them using normal desktop software impossible. So, our claims handlers have had to read through them line by line or listen to calls back in real-time, searching for the crucial piece of information they need to decide whether each customer has an eligible claim or not - the needle in the haystack.
This is time-consuming and, therefore, costly, and it inevitably means customers waiting additional days, weeks, or even months to get a decision on their claim. When we're dealing with pensions and investments, it can be someone's life savings at stake, and we want to make sure we're letting them know if we can pay them compensation as soon as possible.
How does it work?
We have designed and developed the Resolution Data Lake in-house. It uses a combination of optical character recognition (OCR) and search functionality, which means we don't have to manually assess every single page or document in a customer's file.
The tool searches through the data lake and highlights keywords within the documents and flags where they appear. Our expert claims handlers then carefully review the specific pages or files to ensure the findings are accurate. What we’re seeing so far is not only a reduction in the sheer manual handling time for each claim, but also larger reductions in the overall time a customer waits for a decision – as we also spend less time requesting (and chasing) information from third parties.
What are we doing next?
The future of this tool is really exciting for us. As well as improving the customer experience, it's helping us prepare for future failures. At a time when things are more uncertain than ever, it's giving us even more confidence in our ability to handle large, and sometimes unexpected, volumes of claims data at once.
It's also allowing us to begin to proactively analyse and assess the entire data set from a failed firm, without even needing to involve individual customers. This means, for some firms, customers may not even need to make a claim. No application form to complete, no paperwork to find. This also means there's no need to consider using a claims management company either, so no fees to pay - as our service is always free for customers.