The disruption they can't see coming: Insurance that works before Things Go Wrong
Insurance is one of the oldest industries around... but what happens when the data moves over to the customer side?
Hi everyone, thanks for coming back to Customer Futures.
Each week I unpack the disruptive shifts around digital wallets, Personal AI and digital customer relationships.
If you haven’t yet signed up, why not subscribe:
Hi folks 👋
Another deep dive for you. This time about the insurance industry, and what’s coming.
We take a longer look at:
Why insurance is a ‘Future You’
Lots of data about lots of people
Flipping it over to the customer
Is that too much risk, and too complicated?
Disrupting one of the oldest industries around
I’m just not sure the insurance industry can see what’s coming. Can you?
Let’s Go.
In 2013 Angelina Jolie was given a very high probability that she would develop breast cancer. Her medical team discovered she was carrying a rare gene.
And so she took the remarkable and courageous decision to have a double mastectomy.
This was ahead of any actual diagnosis. All based on her personal data and some cancer models.
Right now, a particular risk might be invisible to you, too.
But if it became visible, would you take such a brave decision? Or maybe make positive changes in your daily routines and habits?
My bet is that given the right data - and trusted guidance - most of us would act differently today to protect our tomorrow.
Insurance is a ‘Future You’
When you take a step back, it’s what the insurance industry does. Understanding your risks and being there for you when Things Go Wrong.
But it’s kind of a weird business if you think about it.
They make money by predicting the ‘future you’. And specifically by guessing when, or how frequently, Bad Stuff will happen to you.
Take life insurance.
Yes, it’s about optimising the investments from the premiums you pay. But really, the insurance companies are working out how long you will live. And how much they’ll need to pay out.
Ultimately all types of insurance are about risk. And therefore they’re all about data. To make money, an insurance business needs:
Lots of data - about who you are, where you are and how you live
To be brilliant at predicting things - about investments, but also the chance of you dying, for example
On the first, insurance companies have unfathomable datasets about everything you can imagine.
Lifestyles. Spend patterns. Weather information. Traffic. Hobbies. Even blood results. How often, when and why you interact with them and other organisations.
I once learned that an insurance company will score you based on how and when you renew a policy. Which digital or paper channel? Are you too early, right on time or late? Do you need cover at short notice?
These things become excellent indicators that you will be a ‘good’ or ‘bad’ customer for them (read: profitable).
They’ll also look at where and how you show up in the digital economy. Rating and assessing the datasets they can pull from trusted 3rd parties, and of course each other in shared industry databases.
Some of the more progressive insurance companies even offer tailored - and usually cheaper - insurance packages based on your behaviour.
Collecting data like how frequently you go to the gym, or how carefully you drive. They get your consent to plug into 3rd party data streams - like your gym attendance, or a ‘black box’ you install in your car.
There is a whole other post to be written about the digital- and data-divides of those behavioural approaches. Not to mention the potential bias and privacy implications.
But here I want to talk about the data and the models. How insurance companies can be good at the second thing.
To be brilliant at predictions.
Lots of data about lots of people
Insurance companies have had machine learning and AI for years. Analysing the risks in all that personal data.
What kinds of things happen where I live. Patterns and likely claims from my job type and role. Risks about who I live with and how. My demographics. And the patterns and risks in my biology.
Insurance companies hold very large datasets about very large populations. All so they can understand the chances of Bad Things happening to specific groups and individuals within them.
The insurance companies then overlay these data models together with government and public health databases, environmental information, and even satellite data.
They can then begin to calculate and predict the most amazing things.
The probability of my house flooding or catching fire. Of a car crash happening on my street. Of me getting cancer.
But there’s a theme that we return to over and over again at Customer Futures. That all this modelling - all this value - is on the business side.
These data models are for developing pricing. For managing risk. And for running the business.
Flipping it over to the customer
Today most insurance companies provide generic advice to customers. How to stay healthy, how to look after your home and so on.
But it’s not specific. It’s not personal. And ultimately it looks and smells like marketing. It’s all produced, they’ll say, to foster customer trust.
To show that they know what they are doing (‘expertise trust’) and that the company cares about you and your future (‘motive trust’).
But in Angelina Jolie’s case, it was her medical team that made the prediction. It was her doctors that helped her understand the options, and to make the very difficult decision to operate.
Insurance businesses don’t do that. Instead, they are designed to be there for you when Things Go Wrong.
So rather than just give people advice and guides, why not flip the whole insurance model?
What if insurance companies could hand that data, and selected risk models, over to the customer’s side?
Could people themselves put those models to use? Perhaps these enormous datasets and predictive platforms could become even more useful in the hands of the individual.
Seeing some of your own patterns and probabilities might help you manage your own risks. And like Angelina Jolie, help you take steps to drastically reduce the chances of them happening.
It could make a real difference to people’s lives.
Too much risk, and too complicated?
Surely it would be crazy to give people their own, huge data risk models?
If the customer could only use a spreadsheet to look at the data, then yes, it would be mad. Because spreadsheets are only used by power users like management consultants, people in finance, and data geeks.
You know, the kind of people that use ‘conditional formatting’ for fun (you know who you are).
But soon the individual will be able to put all that data into their Personal AI.
Why?
Because it’s precisely what AI is excellent at. Pulling together messy, often unstructured datasets, and making sense of them. Making predictions.
My Personal AI could, under my control, even sprinkle in all sorts of other personal data. The stuff that the insurance companies can’t - and shouldn’t - get access to. Like my shopping habits and food intake. Like my social network. Like my Strava feed or my real-time heart rate.
Would consumers care though?
If it’s boring and difficult (see spreadsheets), then no. But if it’s easy, fun and helpful, then I think it’s very likely.
In fact, I think potentially this is a massive new market about to open up.
Millions of people already flock to ‘fun’ social media filters and toy apps that can make your face appear older or even younger - all in real-time (the younger option triggers quite an emotional reaction).
These are the just same predictive AIs, but trained on a different dataset: your face, plus some ageing models.
Disrupting one of the oldest industries around
Instead, a Personal AI could combine my personal data with an insurance company’s own (anonymised) risk models.
The result would be a fantastically interesting new opportunity:
To give individuals smarter information about risks in their lives, and options and recommendations to address them.
To disrupt the insurance industry, with a new set of consumer products, new operating models, and new revenue streams.
Because people will probably pay for such data-driven personal insurance insights.
Flipping the insurance data models over to the very customers they describe could open up entirely new business models.
Now, this isn’t about giving insurance companies more data. Because only the individuals would access and hold the data, privately, and then run models using their own AI. We know this is now possible with the latest on-device and ‘Small Language Models’.
But it becomes disruptive to one of the oldest industries around.
Smart insurance companies could pivot from selling ‘here for you when things go wrong’ products to ’before it happens to you’ services.
But remember folks, all this only becomes possible when you start on the customer side.
When you give the data back to people. When you let them model things for themselves. And when we give individuals their own Personal AI.
Understanding ‘future you’ is going to be Big Business. And Personal AI models are going to be central to this new market.
Yes, some insurers are getting close, by developing tools like Aegon’s Retiready. But it’s still all on the business side. It’s Personalised, not Personal (there’s a difference).
Starting on the customer side to create new risk models seems inevitable.
The AI tools are already here. The data sets are available. And it would work within the existing regulatory framework for data processing. All that is missing is the product and commercial innovation to start on the customer side.
Whoever you are, wherever you are, it’s likely you’ll need insurance for something or other.
So we should really start looking into this soon.
To build out these Personal AI tools for personal risks. And we can do it for people and with them. Before it’s done to them.
Your future you will thank you.
And that’s a wrap. Stay tuned for more Customer Futures soon, both here and over at LinkedIn.
And if you’re not yet signed up, why not subscribe:
In my opinion this line of thoughts hits the bulls eye! For the first time I see here the combination of a human (body), a digital identity and the wider idea of describing reality by means of interactions. (watch this to understand : https://www.youtube.com/watch?v=4isz6iQbrYM&t=90s). I really think this is the reason why and how big tech is creating ecosystems and use AI to find the patterns in the logging of actions; by that means they can identify good (user) and bad (hacker; malicious). I even question if the idea of zero-trust isn't simply a orchestrated idea just to get as much "action-logs" as possible, just to bolster the model. I really belief here's an tremendous opportunity; I offer you a direct participation on an initiative to start a company to build such a service!