Processing AI

This week I was lucky enough to attend the SmithNovak NPL Europe conference in London. It was an interesting day of discussion through the debt purchase and servicing processes. The industry, which after what seemed to be seeing a slow down during the pandemic, is clearly now expecting volume to increase.

Feedback from most on the day were expectations that volumes (# of portfolios for sale) will be increasing towards the back end of the year, although the smart comments I felt were pointing to this being slightly later and increasing only in early 2023 (lagged due to the sale process). This will be likely led by Italy and Southern Europe, the UK, with the rest of North-Western Europe then following suit.

Despite the real shock for many economies over the last few years, the sense was this will ultimately probably not be on the same scale as what we saw in 2008. This is in part due to the government support provided, but also interestingly due to the greater awareness and sophistication of creditors, who are pricing and understand the process better too.

Connecting facts

Sometimes at these events, it is also not purely what is said that is interesting, but the connections between the discussion, linkages and reactions that can make you think too. Here are some of my observations.

Artificial Intelligence

AI has frequently been used as a catch-all term for many things from process transformation, data capture, data science, and everything in between. In the AI discussion this was neatly broken down into more discrete blocks I felt.

  1. Make sure you have the right data. Use internal data, scan documents, if you need to
  2. Structure your data. It can take time, but is a critical step
  3. Define your business problem. Businesses are about solving problems for customers and employees. Find areas of pain where you can make improvements, being human-centric is key.
  4. Don’t be afraid to redesign the process. Good process design can lead to better outcomes (and easier modeling/data analysis)
  5. Build your AI models and make sure you have enough processing power, to build and rebuild. AI is a processing power-intensive process
  6. Use data analytics and feedback loops to continue to improve and optimise. It never stops.

All great insight I thought.

Quantum Computing

For the first time, at a work event/conference, I heard the words Quantum Computing used. It was in the context of the processing power needed to run and re-run models.

Processing power is needed and will be increasingly needed, to model with more and more data, doing this faster and repeatedly. (It is a great question to ask any vendor in terms of their capability in this area too).

Quantum computing is of course a potential future solution to enhance computing power.(just like GPUs were in the early days of AI).

The mention of this was a micro-indicator, that this really could be in our future sooner than we think.

COVID culture and masks

With attendees from all over Europe, in the UK, and in a non-restricted COVID environment, our reactions were interesting. Some of us initially had masks on, some quickly discarded and comments of “it’s nice to see everyone’s faces again” were made. I felt it really reflected the mix of local restrictions back in home markets. We are all still adjusting.

Investors vs Servicers

Lastly, sellers, investors, portfolio managers and servicers are of course all part of the NPL chain. They were all represented at the conference.

However, despite being intrinsically linked the culture and focus between the groups came across as really quite different.

On one hand, there was talk of portfolio performance, securitization, ESG, IRR, processes, and techniques to increase returns for investors. It was all very 50,000ft and quite removed from the customer.

Then on the other hand there was a discussion of individual customer situations and the strategies to improve outcomes. It was all much closer to the customer.

There was a cultural gulf between the two perspectives at times I felt.

At the end of the day, all debts do have a customer involved. An originating creditor has lost money, and any recovery still needs to come from the customer, often representing their livelihoods and assets.

We just need to be careful, in the rush to look at returns, not to forget this human element. This is not to change the process per se, maybe more to just make sure we are careful to change the language involved sometimes…

Have a good week everyone

This entry was posted in Opinion. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *