Last week I went to visit a prospective client. As in all such fact-finding meetings, I began by asking the team about the functionality of their current underwriting and exposure management platforms. To understand the process they face when renewing a risk, I ask them to select an expiring policy.

“I just need to pull the latest version of the claims data from Jim’s Excel file,” the most eager of the assembled members said. “It’s easy. We start a new claims spreadsheet for each risk when it first incepts. This one is from 2017. I just have to pull it up.”

Nervous laughter burbled from the others around the table, who realised that this “easy” process was just the type of pain-point that hampers the team’s efficiency.

I went into mollify mode. I realised something long ago, when I was a CUO on the other side of the table. Pointing out the woeful inadequacy of target clients’ incumbent systems, then claiming haughtily how your solution is better, is the fastest way to be shown the exit.

“That’s normal for companies like yours,” I said. “But [insert name of company everyone wants to emulate] has adopted a new system that makes it easier.”

The normalcy part is so embarrassingly true. All too often, syndicates, companies, and MGAs store claims data in one system (typically in a bunch of individual files), and policy data in another. Meanwhile the risk data resides on Jim’s thumb drive (usually with a name like ‘Q2_backup_FINAL_reallyFINAL.xlsx’… which is why the underwriting assistant still can’t find it), and your preferred pricing model is living in Karen’s RStudio session from 3 laptops ago.

But I don’t need to share that detail. They know. Instead I jump to the solution. “[Hero company] now uses a data lake.” The team around the table has heard of this before, years ago, but doesn’t want to admit they are way, way behind the times. So they say nothing.

“Putting a data lake at the centre of your entire IT ecosystem,” I say, wanting to include a sexy buzzword, “provides a single source of truth that lets you get rid of all the spreadsheets.”

The COO pipes up. “But we don’t want the cost and, err, the disruption of replacing everyone’s underwriting and claims platforms,” he says.

“No need!” says I. “You don’t have to have the same system for everyone. You don’t even need all the underwriting or claims teams to operate on the same system. Plus, you don’t need everything in the same format.”

I let this sink in before extolling other benefits.

“Meanwhile, your data is secure, clean, and fit for governance and reporting. You always know who’s been into it. And everyone has access, at least to the parts of the data lake they need. They can always, easily find what they’re looking for, without having to call for help from that stalwart underwriting assistant who doubles as the unofficial IT department.” I nod at the chap who is still unable to locate Jim’s Excel document.

A data lake gives you this:

1. All the data, pooled. No more silos. Everything resides together: policies, claims, risk scores, weather data, IoT from smart homes, PDFs from Swiss Re Sigma, and even the market research that intern did before Covid. At last you can retire all those spreadsheets named FINAL_FINAL_THISONE.csv.
2. A floatie for everyone. Your cyber insurance team can hang onto Python whilst the mariners cling to Excel. And the actuaries? They can still have the 12 dashboards they built in Tableau. Everyone can use their own tool. Same lake. Different floaties.
3. A governance lifeguard. It’s easy to keep your data lake secure with role-based access, encryption, and metadata tags. All changes are tagged, even when people just go in for a dip.
4. A future-proof(ish) utility. The data lake supports innovation. Whether you want to build AI models to predict risk from satellite images, or to ingest fleet telematics directly from individual reinsured lorries, you’re good to go.

“I’VE GOT JIM’S FILE!” the underwriting assistant shouted to the meeting before I got to point five.

“Too late,” said the COO. “We don’t need it any more.”