Antilo – the MGA on a mission to use machine learning to better price risk | IT Interview

No comments

The MGA intends to help underwriters price risk more accurately using technology as well as bridging a “data gap” in telematics

Twelve months ago, insurtech Concirrus partnered with managing general agent (MGA) Antilo.

And after launching its first insurance product for taxis operating in a high-risk area at the beginning of 2020, the product was somewhat hampered by the pandemic.

With fewer people driving during lockdown and subsequently less driving data, the product encountered a challenge.

Nevertheless, Antilo is on wider mission to change the face of data for good, using telematics, machine learning and big data to help underwriters to better price risk, as well as delivering a better offering back to the customer.

This is because current underwriting models within the commercial automotive insurance industry lack a detailed view of changing risks according to the MGA, as premiums are set annually, and based on previous claims and data sets.

Concirrus recently released a report ‘Antilo: Differentiation through innovation in Big Data and Machine Learning’ taking an in depth look at the underwriting model.

The report stated: “This simple snapshot view of an account does not allow an underwriter to understand exposure over time. Insurers are therefore limited in the scope of products they can offer.

“Insurers must evolve their methods of data collection to stay ahead. Ingesting big data, in any format, to derive behavioural trends and actionable insight provide a far more comprehensive view of risk.”

As lockdown unwinds, Antilo intends to use the data from its current insurance offering to inform future products.

Two ends to insurance

The MGA believes that a better understanding of risk could help customers to access products with performance-based benefits and allow underwriters to be more competitive with pricing and product differentiation.

When asked about the proposition, Vivienne Gilroy, managing director and one of the founding members at Antilo, told Insurance Times: “The market moves but some parts of the market don’t. We could see very old [legacy] systems within markets that have traditionally been built and hold a lot of data – to change those becomes a vast investment.

“There are two ends to insurance – the product and the price, and the claims. One is no good without the other being controlled. If we could control claims more efficiently and speedily, this would take time out of the delays, which cost money.”

Gilroy joined Antilo originally as an underwriter and shareholder before taking up her current position as managing director.

She has spent thirty years working in the automotive insurance industry and set up her own brokerage in 1989.

Founded in 2014, Antilo offers commercial insurance to brokers in the UK. It specialises in the niche motor insurance market and predominantly the taxi and commercial vehicle sector.

More data, more problems

Telematics has long been hindered by inconsistent data formats making it unreadable to others, causing problems with operations as well as preventing a shared view of risk in the automotive value chain.

The MGA’s data analytics platform uses machine learning to highlight trends from large data sets and spot behavioural factors that align to risk.

It cited a greater frequency of fraudulent claims and short-term contracts in the industry.

Its new approach pairs historical data to connected products based on real-time data.

Telematics has often been perceived as invasive. When it comes to fraudulent claims, video is valuable in the settlement as it shows exactly what happened.

Gilroy added: “Big data and machine learning allow us to accurately understand the market.

”Automation allows us to apply our knowledge in a manner that improves results. We can identify new segments and develop specific products that better serve the market. Specifically risk areas which we wouldn’t typically operate in.”

And as customers embrace technology more and more, Gilroy said that the MGA can get to know client behaviour instead of just setting rates.

Changes to customer perception and legislation have lowered barriers to entry for the adoption of telematics-led policies.

False positive alerts

Meanwhile, some firms have focused on internal requirements rather than considering the impact on the value chain.

This challenge, Concirrus said, has been exasperated by “the fact that telematics has yet to deliver in terms of its fleet operators”, according to the report.

Telematics devices have historically been triggered by an overwhelming number of false positive alerts relating to first notification of loss (FNOL), and to date insurers have had to manually validate each notification, which can be time consuming.

Antilo’s proposition allows FNOL to be gathered immediately and is triggered by the movement of the vehicle.

“It’s not a short-term fix, it’s something that’s going to be relevant going forward,” Gilroy said. ”But we have always known that if you have to wait for a customer to report a claim, the slightest of delays incur costs.”

Helping brokers

Within the taxi sector, the taxi plating authority suggests to insurers which area a driver has been plated and works in. The broker and insurer would therefore apply this area within a quote and rating offered.

The MGA hopes that by improving claims frequency it can mitigate losses and get better capacity, as well as bridging the ’data gap’ between telematics providers.

Gilroy added: “Through the protection of digital technologies, we can become competitive in higher risk regions. This increase in competition is healthy in the market.”