Lemonade's "A.I. Jim" Claims World Record for Fastest Claim Paid

Brian Anderson

Executive Editor
100+ Post Club
656
Lemonade is claiming to have set a (fictional) “World Record” for the fastest insurance claim ever paid – 3 seconds.

They sent out a blog post detailing the claim (link below), which was completed using the policyholder’s iPhone 7 and the Lemonade app. From the story:

Between 5:49:07 and 5:49:10, A.I. Jim, Lemonade’s claims bot, reviewed Brandon’s claim, cross-referenced it against his policy, ran 18 anti-fraud algorithms on it, approved it, sent wiring instructions to the bank for the transfer of $729 (Brandon has a $250 deductible), and informed Brandon of the good news.

The blog concludes by saying “A.I. Jim” is “still learning” and often escalates claims to a human claims officer. I wonder what percentage of claims get escalated and what percentage are handled solely by the A.I. chatbot?

https://blog.lemonade.com/2017/01/01/lemonade-sets-new-world-record/?utm_source=pr&utm_medium=email&utm_campaign=firstinstantclaim&utm_content=waitlisters
 
Smart strategy on their part. I wouldn't be surprised if they take clear losses on claims just to perpetuate "how quick" they pay claims. You'll have a subculture of dishonest people & fraudsters take advantage of this in no time (especially in NY.)

----------
 
Hmmm, within those 3 seconds, how did they evaluate the value of the claim? Was there already an estimate done for the repairs? If so, then it took a lot more then 3 seconds.

As 1822 said, this is solely about perpetuating a myth about how fast claims are paid, not the reality of a claim.

In fact, this is a morale hazard in the making. Lets see, I'm short on rent, let me file a claim and get paid in even a slowww..... 5 seconds and then I can pay my rent. Of course, most people wont understand this problem, till you've seen it happen a few times in real life and watch people get upset that it might take a few days to get a check. They learn quickly to file the claim at least a week in advance next time ;)

Dan
 
It's fair to assume they've accounted for increased levels of fraud.

But...how tragic would it be if a social media campaign started going around about "educating" consumers about fraud indicators for Lemonade Claims so their claim doesn't incorrectly end up taking "too long". I mean we just want to help consumers you know? If fraudsters use that information for wrongdoings...well...they would never do that because Lemonade donates 10% to charity right?

Again, fraud will destroy this. My bet is they pull from the HO4 market after 2-3 years.
 
AI may or may not be the wave of the future, but for some things it seems to make sense.


A future in which human workers are replaced by machines is about to become a reality at an insurance firm in Japan, where more than 30 employees are being laid off and replaced with an artificial intelligence system that can calculate payouts to policyholders.

The technology will be able to read tens of thousands of medical certificates and factor in the length of hospital stays, medical histories and any surgical procedures before calculating payouts, according to the Mainichi Shimbun.

https://www.theguardian.com/technol...-intelligence-ai-fukoku-mutual-life-insurance


FWIW, humans do not currently calculate claim payments. That is handled by computers as it has been for quite some time. Humans review the results and decide to pay or pend for further review.
 
It's fair to assume they've accounted for increased levels of fraud.

But...how tragic would it be if a social media campaign started going around about "educating" consumers about fraud indicators for Lemonade Claims so their claim doesn't incorrectly end up taking "too long". I mean we just want to help consumers you know? If fraudsters use that information for wrongdoings...well...they would never do that because Lemonade donates 10% to charity right?

Again, fraud will destroy this. My bet is they pull from the HO4 market after 2-3 years.


In another post, I believe Lemonade claims that 38% of premiums go to pay for fraudulent claims. Somehow they think their algorithms are going to eliminate that, along with attracting only honest people because of the charity donation aspect and, of course, their signing an honesty pledge (what the rest of us call a proof of loss).

I'm curious as to what happens when they deny a claim based on these algorithms. Are the algorithms revealed to regulators? What are they based on and, if used to deny a claim, are they in compliance with existing Fair Claim Settlement statutes? If you examine most of these statutes, they have very rigorous REQUIREMENTS for ascertaining the validity and value of claims.
 
Lemonade released more details from its first 100 days (see link below). Among them:

• 25% of people who get a quote end up buying
• 90% of their customers buy renters coverage; 10% buy homeowners so far
• 35% of their premium comes from Manhattan
• 6: Number of claims filed in company’s first 100 days
• 12%: expected final loss ratio for 2016

Insurance Forums | Lemonade
 
We had a tornado in Albany, GA on 1/2/17 and then another on 1/22/17. Four people killed. Haven't had a tornado here in 40 years. Millions and millions of dollars in damage. I wonder how quick lemonade would pay these claims. And they would probably be in receivership before all the claims got paid.
 
Back
Top