AI will almost certainly become part of this as well. The Bank of England says the use of AI in credit risk management is still relatively early, although it is already being used at various stages of lending and expects more advanced AI increasingly to inform areas such as credit underwriting.

There is plenty within an underwriting process that technology should be able to make quicker. Gathering and analysing information, reviewing documents and data, identifying patterns and dealing with some of the more repetitive work all look obvious areas. For boards looking for a return on their technology investment, credit is therefore an interesting place to look.

I don't think that necessarily leads to fewer good underwriters. It could change what firms value in them.

If technology does more of the processing, there is greater emphasis on what happens after the information has been produced. Someone still needs to understand the borrower, question assumptions, consider the downside, structure the transaction and ultimately decide whether the firm should put its money at risk. In specialist lending particularly, the circumstances that make a transaction interesting are often the same circumstances that mean it doesn't fit neatly into a standard model.

There is also a longer-term talent issue which I think is worth considering. Today's experienced Credit Directors and senior underwriters developed their judgement by looking at an enormous number of transactions earlier in their careers. They learnt from straightforward deals as well as difficult ones, and from seeing what happened to those decisions several years later. If AI gradually removes more of that work from junior roles, firms will need to think carefully about how the next generation gets enough exposure to develop the same judgement.

That could have quite an impact on the structure of credit teams. Technical underwriting knowledge will remain important, but so will commercial judgement, sector knowledge, structuring ability and the confidence to challenge a decision. Firms will also need people who understand what technology can do without assuming that its output removes the need for experienced oversight.

For CEOs and boards, I think this becomes part of a broader question about what they want their lending business to look like over the next few years. If volumes increased significantly, moved into different sectors or came through new distribution channels, would the current credit team have the capacity and experience to support it? Equally, if AI materially improves productivity, are the skills that were valuable in a credit function five years ago necessarily the same ones the firm will pay a premium for five years from now?

We already spend a lot of time speaking to experienced underwriters, Credit Directors and senior credit professionals across specialist lending and asset finance, and I expect those conversations to become increasingly important as the market changes.

The next cycle may require fewer assumptions about what a credit team has traditionally looked like and more thought about what the business actually needs from it. For some firms that will mean technology, for others additional underwriting capacity or different sector expertise, and for some it will be about strengthening commercial judgement at a senior level.

Whatever happens to rates, I find it difficult to see experienced people who know how to lend money sensibly becoming less valuable.