Guillaume Bonnissent’s Insurance Technology Diary: Coal or horse?

About 15 years ago my wife and I decided to rip out our family bathroom and start again. It was so dated that Thomas Crapper himself may have installed the fixtures. There was just one problem: we couldn’t find a plumber.

Correction: we couldn’t find a qualified plumber who was available. A couple of them said they could do the job in about a year’s time, and would pencil us in, provided we paid a deposit, which didn’t match my understanding of the time value of money.

Flash forward about five years. The number of plumbers seemed to have multiplied, so availability was no longer so much of an issue. However, since the first round it had become clear that our boiler needed replacing. As anyone who’s had any major work done on their home will know, doing two jobs at the same time brings less than half the grief of doing them separately (since the latter yields two thick layers of dust to clear and at least one sustained session of ‘OMG-not-this-again’ despair).

Unfortunately few of the available plumbers appeared to have the necessary qualifications. Since many have been trained out of the country (like me), the available plumbers didn’t tend to have the national gas certification standard, which didn’t match my understanding of the wisdom of health and safety regulations.

Another five years, and we tried again. All the foreign plumbers had moved home (presumably since their eligibility to work in the UK had been voted away). I was only able to secure a new boiler and a new bathroom when I met the father of a friend of my youngest son at a birthday party we threw, learned that he was a plumber, and in effect wouldn’t let him go home until the job was as good as done.

Plumbers are coal. Until blowtorch-fitted human-esque robots are able to park their van awkwardly in front of the neighbour’s house, come into yours with their boots on and perform plumbing services powered by the wisdom of artificial intelligence, the job of plumber will not be usurped by AI.

I read a marvellous article by Annie Lowrey in The Atlantic before Easter. It’s titled “How to Guess If Your Job Will Exist in Five Years”. It’s the big question posed: will AI eat my lunch? Lowrey says workers must choose if they are horses (26,493,000 on U.S. farms in 1915, versus circa 700,000 in 2015), or coal (the global use of which has escalated dramatically since 1915 and grown almost every year since).

The music begins when Lowrey compares the decline of the U.S. agricultural employment from more than half of the workforce to about 2%. She cites also the occupations drayman, telephone operator, woodchopper, hoistman. Whilst acknowledging transitions were drawn out and painful, humans have in general managed the tides of change better than horses, she says.

“When the combine (harvester) rolled onto the alfalfa field, horses did not see the writing on the barn wall and start applying for factory jobs. They didn’t learn to code or attend community college. They stood there and ate carrots.”

She also showed that the impacts of technology – and of life and human need – may not be as initially anticipated. Instead of making radiologists obsolete, for example, technology created a shortage of them. Meanwhile, the most common job around San Francisco isn’t in AI, it’s home care.

For insurance employees, it’s not rocket science. If your job involves information sifting, parsing, researching, gathering, reformatting, or any of a number of other low-value-added tasks that we already know can be done by AI, then that part of your job is likely to go. Probably part of your team.

Ditto, ultimately, for the functions in your job where the value you add is standardised knowledge that can be easily learned from books or YouTube videos. So there go the wording experts, a whole lot of the modellers and a great many of the lawyers. Farewell. Down the line (and maybe already), AI can probably do a better job of assessing your inwards portfolio and structuring an outwards tower that optimises your reinsurance spend to match your risk appetite exactly.

It’s essential, therefore, to find the coal-like element of your role, and move towards it. AI can help you structure a new product (we’re building a platform to do that right now), but it’s not so good at spotting a valid market niche and pairing it with a distribution partner. It may be able to structure a perfect set of reinsurance program options, but they still need to be sold to the cedant by someone who can also take them out to lunch.

Consumers in the UK may already be willing to outsource their insurance-buying to an AI broker aggregator that appears in the form of a meerkat or an opera singer, but (so far at least) corporate insurance buyers want better service. Carriers want humans to do engineering inspections. The judgement of a few decades’ experience cannot yet be coded. As Lowrey says, “At least for now, artificial intelligence is changing how doctors do their job, not eating their lunch.”

Still, I have advised my sons to consider a career in plumbing. No matter what, they’ll be in demand.

* Like every Insurance Technology Diary entry about AI, this one is accurate only to the best of my knowledge at the time of writing. The pace of AI progress is so great that I cannot guarantee it remains so now that it’s finished, let alone when you read it.

Guillaume Bonnissent is CEO of Quotech.