Last week the London Market Group launched consultation on the structure and fields necessary to extend its Core Data Record project to claims. I am not sure the CDR is being used yet by anyone for anything, but already we’re on to CDR v3.2, so we’re making progress. But do we need structures and fields?

The consultation is packed with deceptively simple-looking questions, such as “Do you think any additional fields are required for claims process? (Yes/No).” That’s the sort of question that can be answered only after rather a lot of use in anger, which is the only way you’ll realise you regularly need a field which, in the abstraction of a consultation, you couldn’t pull from the air.

That in turn speaks to data richness, which I will get to. In the meantime, the CDR. To recap, according to Lloyd’s, the CDR “holds the critical transactional data required to be captured at written line [level?] across four downstream processes: premium validation and settlement, claims matching at first notification of loss, tax validation and reporting, and regulatory validation and core reporting”.

Lloyd’s says: “Data is either captured in the MRC v3 during risk placement and easily extracted to complete the CDR, or – for those using a data-first approach – the CDR can be used to help create the MRC v3.”

It’s all so wonderfully easy.  Data is “captured … and easily extracted,” or, more loosely, “used.” What this paragraph fails to point out is that each of these processes requires a compliant system built to capture, extract, and use the data to be delivered into or extracted from the CDR. Not so simple, then. Nor cheap.

And there’s another complication. “The CDR aligns with ACORD technical standards, which are the agreed method for submitting and exchanging data within the London market.” Unfortunately ACORD standards (unlike OASIS) are not free. Unfortunately ACORD will not even openly tell you how much they cost, even when you ask directly. (So they must be a lot, right? If you have to ask, you cannot afford it.)

Buying into ACORD is the table stakes for using the CDR. That’s what comes of having someone from ACORD (rather than, say, someone from free-to-use OASIS) on the London Market Group’s wise-sounding Data Council. Perhaps that’s why in June the CEOs of ACORD, the IUA, LIIBA, Lloyd’s, the LMA, the LMG, PPL, Velonetic, and Verisk (a collection of leading QUANMOs, or ‘quasi-autonomous non-market organisations’) issued a widely publicised “open letter to the insurance industry” that said: “Now is the time to take action. All insurers and brokers who wish to take full advantage of a coming digital market will need their data to align to the ACORD GRLC Standard.”

(Perhaps, I humbly suggest, if the signatory bodies built something that the market wanted, and that they didn’t have to spend enormous additional sums to use, then their CEOs wouldn’t have to beg the market to adopt the system. But what do I know?)

Back to the consultation. I am amused by the idea of being asked to fill in an Excel spreadsheet to comment on the structure of reports that will be made on spreadsheets. I wonder how many people will complete theirs with ChatGPT.

I am dismayed, though. Whether it’s called a core data record or a spine or a bordereaux, it’s a consultation on a new spreadsheet. Instead we should use the market’s collective intellectual and experiential chutzpah, plus a sprinkling of AI, to rethink the process entirely. We should figure out how to use currently available technologies to eliminate spreadsheets altogether. Then we won’t need data standards.

At the very least, the LMG should ask their favoured Large Language Model how the market can leap ahead to the next level, the one where spreadsheets are yesterday’s technology, and all data is made compatible on the fly.

Amazing technologies are already in use by London-market companies that deploy AI to take unstructured data and make it compliant with the square holes in spreadsheets. But by getting rid of the spreadsheets, we can remove this step entirely from the process. We simply need to think creatively.

Instead, the LMG is asking market companies to spend fortunes building new systems to work with old spreadsheets.

“What’s wrong with them?” I hear you cry. “We’ve always used spreadsheets.”

They have a fundamental drawback. They rely on data standards, which are the antithesis of rich data. By forcing everything to fit into a well-defined set of fields, anything that doesn’t fit is abandoned. AI could be used to help address this. At the very least, it could be deployed to help us make standards looser and more flexible, whilst retaining their benefit of making data is easily interpretable.

Spreadsheets are made up of square holes. Data very often comprises round pegs. AI is the tool that can make everything fit, without paring it down. Let’s have a consultation about that!