How AI helped automated car insurer Naked Insurance take on home cover too

Ernest North, Naked Insurance (Supplied)
Ernest North, Naked Insurance (Supplied)

Naked Insurance, the automated insurer famous for allowing consumers to pause their cover, is branching into home insurance, moving a step closer to competing with short-term traditional insurers like Outsurance and Santam.

The artificial intelligence (AI) driven insurer, which broke into the scene in April 2018, is backed by Hollard and Yellowwood, who also invested in automated life insurer, Simply, late in 2018. Naked spent the past 20 months selling only motor insurance. Big short-term insurers on the other hand cover a range of possessions, from car to boats, home, valuables, jewellery and businesses.

That Naked is adding structural, home contents and valuables in its offering makes it the first among its AI-driven peers to move towards becoming a comprehensive short-term insurer. For instance, Granadilla and Pineapple only offer valuables insurance, although the latter plans to launch motor cover soon. Other fintech players operate in niche markets, like InvestSure, which covers investors against share price losses or have chosen to go the broking and distribution route, like Ctrl Insurance and Click2Sure.

Thirsty for change

“We are consumers too, and one of the things that happens when you phone to cancel your cover, they offer to drop the price. That on its own is admission that the price you’ve been paying could have been lower. So, what we said was; it was time for insurance for home insurance to enter a new era of business too,” says co-founder, Ernest North.

The selling of home cover, which the company has tested on select customers for few months now, is also completely automated. Customers buy it through the app, adding and adjusting items on their shopping baskets, as they do with any online purchase. The AI bots estimate, given the person’s profile, age, the area they live in and their credit records, how much cover they need. But customers have the option to adjust the bots’ estimates manually if they know values of their building and other items. The only proof the company requires is a “home selfie” and snaps of their valuables.

“The model is built in-house,” said North. “There are no humans involved in the buying process. The technology is 100% self-service. Every day that this thing is live becomes more accurate, because we are storing all the data every time someone does a quote to see how far off our estimates were to the actual value when people claim.”

How AI-driven insurers keep costs low

Most insurers in SA now enable customers to buy online, whether through their websites or apps. But even if the purchase is made online, there are back-end systems which consist of human beings that reconcile data captured by the website with actuaries’ calculations and file these for the finance teams.  

AI insurers, on the other hand, are almost completely automated. One area where they still struggle, said North, is in the process of processing complicated and slightly complicated claims. He said at this point less than 20% of Naked Insurance’s motor claims are processed completely by the system. The company still sends a human assessor in most cases to value the damage. But the bots are learning from the data humans constantly feed the system with the goal to gradually process more claims on their own.

“We have cut out, almost all of the costs which for normal insurers make up a big portion of the premiums,” said North.

Springboarding off motor success

Since it is not a public company, Naked did not reveal how many customers it now has in its books. But North said while they initially expected the demographic to be skewed towards younger generation, Naked had “quite a lot” of customers falling out of the 25-35-year age range. He said their claims loss ratio – the portion of the premiums that is paid back to customers through claims sat neatly where their team of 22 developers and data analysts had predicted: between 70% and 80% in the past year-an-a-half.

“Our actual sales volumes are ahead of what we predicted,” said North.

Brent Crude
All Share
Top 40
Financial 15
Industrial 25
Resource 10
All JSE data delayed by at least 15 minutes morningstar logo
Company Snapshot
Voting Booth
Do you think it was a good idea for the government to approach the IMF for a $4.3 billion loan to fight Covid-19?
Please select an option Oops! Something went wrong, please try again later.
Yes. We need the money.
11% - 974 votes
It depends on how the funds are used.
74% - 6472 votes
No. We should have gotten the loan elsewhere.
15% - 1341 votes