In September of last year, McKinsey & Company published a report called, How top tech trends will transform insurance. This excerpt about claims technology caught our eye:

Technology trends have the potential to materially change some of the underlying inputs of insurance products and core functions … [allowing] carriers to more effectively manage risk and make use of complex customer data—a critical step in evolving to a “predict and prevent” model of insurance where data is shared more frequently between parties.

It’s actually quite amazing to think about how far we’ve come from paper applications and telephones as means of submitting claims and how quickly it happened. We’re not quite sure what McKinsey means by will transform, since software has been transforming insurance — particularly in claims — by helping companies of all sorts to submit claims, to manage risk, to analyze customer data, to identify risk trends, and to mitigate the risks posed by those trends.

But that’s okay. That’s not the part that really got us.

Here’s the Head Slap

Later in the report, McKinsey wrote this:

The role of insurers may shift from claims to prevention, whereby they are best placed to identify and reduce risk by partnering with clients and using technology.

Whoa! It can be argued (as we did above) that software is already enabling insurers to play a more active role in mitigating risks. But actually shifting roles from claims indemnification to claims prevention? That’s a whole different kettle of fish.

McKinsey may prove to be correct. But for the moment, we’re not sure how software could predict the occurrence of risks well enough to preclude them. If we take into account the vagaries of human nature alone, it seems like a stretch. How, for example, in any given scenario, can we know Person A will do this, instead of that? Under what specific conditions is he likely to do one and not the other? How can we know?

We’re software developers. We believe in technology, of course. But we don’t want to get ahead of ourselves in predicting the omniscience of software, even the software we develop.

That doesn’t mean we sell ourselves short. But it does mean we wonder how high tech can go.