Customer-ization of Risk Analytics - Data Science, Data, AI, and Cloud
And so, you get two scoops…enjoy! [my favored flavor is as mentioned Pistachio – front and center]
Source: I took this photo (Author of this Blog: Marty Ellingsworth).
But from here, I go forward in this blog to talk about dots, and dots on a map, especially dots on a map sprinkled around over time, and how to use observational data for risk-based pricing. See my latest report here - "Risk Personalization in Auto Insurance – The Rise of the Machines".
The essence of mobility and personalization of risk pricing is how these and other dots interact - including commercial and personal lines.
Everywhere that sensors and people go, the work, home, and life routines of people and machines alike create patterns within patterns that make many things detectable, predictable, and actionable – often before a loss, so prevention, safety, and trusted risk advisory value can occur.
Like the colors of gelato, we can uniquely paint all the connected and connectable devices in our homes, cars, phones, communities, and workplaces. Many upon many more connectable devices and sensors are inside each and let’s not forget larger machinery, and smaller wearables or even implantables. We can paint all those their own colors too.
If you don’t have a sensor on your home, car, or phone, don’t worry there are ways to infer the appropriate colors, or even to enable them to talk and paint themselves.
Machines talking to machines create invisible yet authentic connections across our risk landscapes.
In a simple analysis, mobility starts in reference to stationarity (where you typically sleep at night).
A blue dot for your phone, a green dot for your home, and an orange dot for your car. You are in your “most protected state” when all three dots overlap – you at home with your car in the garage. Construction is known, Occupancy affirmed, Protection at highest level, Environment observed. And mileage zero. It’s a twofer – COPE and Car in a Cage. Only really a Major CAT to worry about.
Risk changes when the dots change, up or down, and then back again to the “most protected state”.
When you drive on a trip (chore, shopping, dining, school, community, work, etc.) the green dot is the same, but the blue and orange dots move away from it together. That is until you park and walk away – then the orange dot stays, and the blue dot moves alone. I did say simple analysis.
For the most part, humans are creatures of habit, and the patterns of any household living situation will dance like a spirograph (a graph making art toy of my youth) where each member in the living situation in a household spins in its own orbit and can interact with all others. Time plays forward.
Every house/dwelling/property may contain a household (or be vacant). Every household may have zero, one, or more phones and zero, one, or more cars that may be in use by one or more adults with or without children, and sure, pets too. Even vacant properties can have active sensors.
Keeping it simple still - for a single adult living alone in one house with one car and one phone, the graphs are easy to observe, and the trips and trip purpose combinations are simple to compose and understand.
In any given timeframe, we can observe “at home”, “driving”, and “on the move” or “busy/waiting”. By partitioning and clustering an accumulation of observations over time on a risk map, we can calculate the sum total of a variety of intrinsic, operational, behavioural, and external risks that when covered by a set of insurance products/endorsements would generate a custom price, perhaps with some terms and conditions tacked on (the most risky may get the most T&C).
During COVID-19, mobility restrictions both enforced and personally elected often stick the orange dot and the green dot together more than ever before (car is at home). The blue dot too (that’s you). Even as a ‘homebody’ at their “home-as-an-office”, the blue dot is mostly at the green dot too, until going out for a walk or run (bike ride maybe). Lot’s of pet adopting going on right now.
The sum-total of being at the “most protected state” means no car crashes, no home break ins, no slips or falls “at work” (fuzzy still here), quicker noticing and mitigation of any home peril (fire, water, smoke, theft), and the potential to mitigate weather (wind driven rain, freezing, etc). And with like changes, the “office-as-an-office” commercial exposures are observably different too.
Arguably the warning “some assembly required” is a fair criticism now, but the brilliant pebbles, or point solutions of insurtech, are collectively massing to end-to-end ecosystem solutions. Text, sounds, temperature, motion, images, infrared, status pings, machinery monitors, security systems, smoke detectors, chemical sensors, freeze warnings, magnetometers, crash detection, and more.
These enablers, effectors, and evidenced-based solvers are now pre-underwriting pre-packaged data assets at the digital twin equivalent to most products in market short of complex commercial accounts or bespoke artifacts. With a smartphone reading of a bar code, you can get a bundled quote in seconds. That same phone can let you report your car odometer, opt in for mobility measurement, or even pair with other smart things to prove how you should be risk rated as well as continuously risk assessed, variably covered, and even dynamically re-priced.
Once observed, you may benefit from getting valuable tips on how to improve your safety, or costs of insurance, or even to opt into deeper engagement with discount opportunities or rewards. Easy, immediate, and trustworthy is the way many customers expect the insurance experience to be now.
If you are a ‘non-subscriber’ to Celent, I am sorry you bounce from my paywall, but pick any paper you like (list of recent reports here Recent Reports on Data, AI, and Cloud) and we can discuss over Zoom where I can walk you through the paper and exhibits, answer questions, give additional insights, or make value added introductions (ping me at firstname.lastname@example.org).