Incremental disruption in insurance industry “like death by a thousand cuts”: Insurance Analytics Canada speaker

The insurance industry has been very incremental in dealing with disruption and needs a drastic shift from historical analysis to predictive, future risk models, a speaker suggested on Wednesday at the Insurance Analytics Canada Summit.

“You cannot disrupt with incrementalism,” Pranav Pasricha, chief executive officer at Intellect SEEC said at conference, held at the Toronto Marriott Downtown Eaton Centre Hotel on Wednesday and Thursday. “Our industry so far has been very incremental. You can’t win in this world of insurtech by doing a thousand small, little insurtech projects. That’s going to be like death by a thousand cuts.”

Pasricha pointed to certain S-class car models from the 1970s and 1980s to today, noting that the industry has not shifted. “There’s still gas guzzlers; they still require maintenance,” he said during his presentation on Wednesday titled Big Data & AI: Time to Unlearn All You Know About Traditional Analytics. “In comes Tesla. What we’ve done is achieved a platform that can auto update, never break down, there was no depreciation, cheaper than a Mercedes, faster than a Ferrari, no emissions, no gas, etc. etc. That’s disruption, that’s paradigm shift.”

The biggest challenge for the industry is technology, particularly outdated core policy admin systems, Pasricha said. “We need to move from structured little data to big data,” he argued. “All policy admin systems are based on the same information that was brought down into the industry in the 1970s.”

Data models are based on historical statistics, with the industry working on the same static data for decades. “We’ve always had this concept of static data models,” Pasricha said. “They do not work anymore. The model for now is big data: you have 10,000 indicators, not 70 indicators of underwriting and they’ll change every single day. That is what takes you from a reactive, backward-looking model to a truly predictive future-looking model.”

To get to a future-looking risk model, insurers should take their information systems from what Pasricha calls “structured, little data” to “true big data.” He suggested insurers run their processes and current actuarial ratings, and then run machine learning systems in parallel “so it becomes a supplementary tool. Here is how you can make a better decision and this is how you should change your coverage and pricing.”

Next, as these changes take hold, the machine learning algorithms work in the background, starting the supplementary rating and underwriting model. Finally, “the machine learning algorithms can now perform for two or three years. That’s where you can take your hands off the steering wheel and now this car can drive on its own.”

In the next two years, Pasricha predicted that “range of gadgets that will monitor every aspect of your life is going to exponentially increase.” In the insurance space, “we are already talking about genetics – how do you get genetics as a predictor of future mortality into life underwriting,” he said. In the P&C space, Pasricha said there are pilots of Internet of Things (IoT) and augmented reality devices on cellphones for industrial safety.

“What we are doing is putting together a whole network of beacons and IoT devices out in industrial sites or office buildings so you monitor the whole work environment and you can monitor a person, a worker, whether they are an office worker or a blue collar worker,” he explained. The devices can monitor the safety environment, hazards workers walk through, safety practices, “do they bend properly, do they lift properly, are they operating equipment properly?”

Also in the next two years, 1,000 different data sources are going to become 10,000 “and the quality and the data feed you are going to get is going to keep changing every day,” Pasricha said. “This is something that the insurance industry, like most other industries, have no concept of. Fundamentally change your data process; change your thinking process. I think the industry is broken to a large extent and we are sitting at an historic crossroads on how to reform it.”

More coverage of the Insurance Analytics Canada Summit

No need for ‘massive upfront investment’ to get data analytics to work: Insurance Analytics summit speaker

Mobile telematics data can offer rich insights into distracted driving and impact loss trends, Insurance Analytics Canada speaker suggests