Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
INSURANCE
INDUSTRY
26 REAL-WORLD
USE CASES
1
TABLE OF CONTENTS
3 REDUCING FRAUD
4 INCREASING SETTLEMENT SPEED
4 BETTER STRATEGIC PLANNING
4 CREATE A MORE EFFICIENT TECHNOLOGY INFRASTRUCTURE
5 CUSTOMER INFORMATION AND RISK ASSESSMENT
5 BEHAVIOUR-BASED PRICING
6 CUSTOMIZED PLANS
6 IMPROVING THE DIGITAL EXPERIENCE
6 CHATBOTS AND CUSTOMER EXPERIENCE
7 SPEEDING UP THE UNDERWRITING PROCESS
7 AVOIDING HEALTH INSURANCE FINES
8 KEEPING UP WITH UPCOMING TECHNOLOGIES
8 FACIAL RECOGNITION
8 THE RISE OF INSURTECH COMPANIES
9 ENHANCED MARKETING EFFORTS
9 FIXING HEALTH CARE DATA FRAGMENTATION
10 IMPROVING HEALTH OUTCOMES
10 QUANTIFY CYBER RISK EXPOSURES
11 UNDERSTANDING HUMAN VOICE
11 NATURAL LANGUAGE PROCESSING
11 NATURAL LANGUAGE GENERATION
12 GAMIFICATION AND LIFE INSURANCE
12 DETECTING OBJECTS WITH COMPUTER VISION
13 PREDICTING EVENTSON A GLOBAL SCALE
13 BEHAVIORAL BIOMETRICS
13 CONTEXT TELEMATICS
14 CONCLUSION
14 SOURCES
The insurance industry is built around risk and the ability of the insurer to correctly
estimate and predict it. By collecting all the information available about an individual or
organization, an insurance company must make an accurate prediction in order to earn a
profit. If this prediction is not accurate enough, instead, the result would be a (potentially
huge) financial loss. It’s not surprising that the insurance industry always uses the most
innovative technologies to perfect its ability to assess customers and optimize their
products (i.e., their insurance plans).
AI is one of these technologies, and many insurance companies are already using it (or
plan to do so in the near future) to get an edge over the competition. AI can help insurers
in a lot of different ways, and it is going to be a must-have tech very soon. So let’s have a
look at what the future of insurance holds and how the innovation brought by AI can be
used to hit the mark even more precisely.
1 REDUCING FRAUD
No matter how precisely an insurer may have calculated risk,
scoundrels can always find a way to steal money from companies
by using fraud. With an estimated cost of $80 billion per year, fraud
accounts for 5-10 percent of claims costs for North American insurers.
Unsurprisingly, most insurance companies rate one of their key
priorities as upgrading analytics to detect fraud before a claim is paid. The problem is
that in such a hectic environment, no human analyst could even gather all the necessary
information about a policyholder in a timely enough fashion.
But what about a machine? An intelligent one can process an absurdly high amount of
apparently menial information to detect even the slightest hint that may indicate that
fraud is occurring. AI-powered identity verification services such as Shufti Pro, IDnow
or Jumio can also help establish the authenticity of a submitted claim by checking the
identity document, insurance card and other related documents. As the chance to detect
unlawful or mischievous behaviors increases, insurers may lower their premiums since
they do not need to account for these losses anymore.
3
2 INCREASING SETTLEMENT SPEED
The same reasoning described in the above point about reducing fraud
can be applied to all settlements, including the (many) legit ones. AI
can process enormous amounts of data in no time, speeding up the
entire insurance machine, and making the whole process of settling
a claim much more efficient. It can be so fast as to settle a claim in
less than 3 seconds, as Jim, the AI used by the insurer Lemonade, showed us in 2017.
Machine learning allows AI to grow better over time by learning from past mistakes,
further reducing the overall effort in the long term.
4
it to the other specialists. When humans need to talk to and understand each other,
integrating information, insights and technology becomes a chore. But if this same task is
done by software-based minds, it becomes much simpler.
With no need for any human resources except for a handful of AI experts and data
analysts, machines can consume enormous amounts of data to produce actionable
insights that can be used to have a clearer understanding of risk at a reduced cost.
6 BEHAVIOUR-BASED PRICING
Together with the implementation of sensors and IoT, AI can help
insurers draw a clearer picture of individual customers and their
behavior, to customize plans around them. For example, sensors
installed in a vehicle may be used to gather information on how the
insured drives – how quickly he accelerates and brakes, how frequently
he exceeds the speed limit, etcetera. This information could be used to comprehend that
user’s behavior, correctly assess the risk (such as how likely he is to cause an accident),
and provide him with unique premium discounts (or penalties) based on his performance.
5
7 CUSTOMIZED PLANS
Modern marketing is all about providing customers with highly
customized products which are tailored to fit their unique needs. Most
customers are more than willing to share their personal information to
receive a benefit in return – such as in the form of a plan with reduced
costs or increased coverage. Machine learning opens a plethora of
new possibilities – its notorious digital agility allows for a quicker integration of data
and variables to increase the individual value of the products offered to customers. Fully
customizable plans that take account of a customer’s budget, habits and lifestyle could
be presented with greater ease and flexibility as experimented with by the Italian company
Allianz1.
6
The cost of running online interactions between the company and its customers will be
dramatically decreased, increasing profits and giving ampler space to insurers to lower
the price of premiums.
7
12 K EEPING UP WITH UPCOMING TECHNOLOGIES
New technologies are going to revolutionize the entire world – self-
driving cars are a prime example of the immense transformations that
we can foresee in our near future. A fully automated world requires a
different approach, and insurance is no different. For example, when
cars will be able to “make decisions” about driving, a completely new
way to determine who is at fault during an accident needs to be established. AI represents
a solution, since it will be able to integrate all the information coming from these highly
connected vehicles and all the technologies surrounding them (sensors, smart poles,
satellite views, smart roads, etc.).
13 FACIAL RECOGNITION
One way to increase customer onboarding is to provide them with
much quicker service. Instead of requiring long, expensive and
somewhat uncomfortable medical examinations, for example, the
startup life insurance company Lapetus allows enrollees to buy
insurance with just a selfie. Their technology uses AI-powered facial
recognition to rapidly assess a human’s lifespan and negative habits such as smoking
cigarettes, and assign risk scores in just a few seconds. And the best part is that facial
recognition technologies are still far from achieving their true potential, so this method is
likely to become even more accurate as time goes on.
8
going to revolutionize the 300-year-old competitor landscape, forcing many of the older
players to reinvent themselves. AI can simply become a powerful force to drive marketing
efforts and show a brand’s continuous commitment to innovation and modernization.
9
17 IMPROVING HEALTH OUTCOMES
AI could be used to fill gaps in care and improve health outcomes so
that the number of hospitalizations and emergency visits is reduced.
For example, the insurtech company Clover launched an AI-powered,
in-home primary care program that integrates genomic testing with
clinical support tools to improve the quality of life of enrollees. The
most vulnerable members receive uniquely tailored plans to reduce adverse events,
assist them with medication therapy management, and provide them with much-needed
preventive care advice. The expenses to deal with some vulnerable populations such
as patients affected by type-2 diabetes, for example, may become particularly hefty.
However, these costs could be cut significantly when these conditions are managed
appropriately, with obvious financial benefits for all parties involved.
10
19 UNDERSTANDING HUMAN VOICE
Some AIs are able to understand the human voice and recognize
words to transcribe them, as well as index and tag spoken words from
audio files. But some of the smartest ones can do more than just
transcribe a few short pieces of text, and can even detect emotions
and wellbeing from the intonation in the voice. One such example is
Beyond Verbal, a smart voice-detection software that can identify physiological markers
through the voice that spot many health-related issues. Interestingly enough, this software
crosses any language barrier. Information is gleaned from tones which come from the
limbic system and is therefore consistent across all languages.
11
amounts of content written in real time, to quickly answering customer questions, or
providing instant feedback for usage-based insurance. Companies such as SafeDrive
have even used it provide textual feedback and driving tips as a part of a behavior change
support system for automobile drivers.
12
24 PREDICTING EVENTS ON A GLOBAL SCALE
Some AI is so advanced that it is able to predict future events on a
global scale, such as climate-related risk (wildfires, flooding, storms)
on time frames of up to several decades in the future. These amazing
technologies such as the one invented by Jupiter use satellite and
sensor data for climate analysis to predict changing climate effects,
create reliable flood maps, and make other forecasts that could help insurance companies
assess catastrophic risk at an amazingly granular level.
25 BEHAVIORAL BIOMETRICS
As the insurance world shifts towards online experiences, fraud has
become “digital” as well. Legit customers’ accounts could be stolen, or
fraudulent accounts could be easily created with stolen data, starting
a vicious cycle of false claims and diverted payments that is extremely
hard to stop and even harder to detect. AI can use behavioral
biometrics techniques to analyze a user’s behavior – from creating an account page to
the time required to enter basic credentials – to spot bots and fraudsters by flagging any
suspicious behavior.
26 CONTEXT TELEMATICS
Contextualized telematics (CT) is different from standard telematics.
CT overlays traffic, terrain, weather, and handling of a cell phone while
in motion to provide a better understanding of driver behaviors. It
can be incredibly useful when paired with commercial fleets since
underwriters can “experience rate” fleets and adjust pricing according
to past data. As Andy Gastley - practice leader from A. G. Roth explained, “To the extent
a business owner can control claims through behavior monitoring and tracking, the
business owner can improve their claims – thus reduce their auto fleet insurance cost.
The CT is the best driver ‘babysitter’ a business can use.”
13
CONCLUSION
Machine learning-based technologies have already begun to affect the world of insurance
by making it more accessible and efficient than ever. AI in insurance is not a disruptive
technology – on the contrary, those who will adopt it as early as possible will be able to
provide their customers with innovative and streamlined products that will harness the full
potential that cognitive learning has to offer.
The impeccable logic and imperturbable reliability of machines can be used to calculate
any risk with absolute precision, reducing waste and making all products much more
affordable. Both consumers and insurers are going to enjoy the benefits of the widespread
implementation of this latest advancement of the era of digitalization.
SOURCES:
https://www.mckinsey.com/industries/financial-services/our-insights/insurance-2030-
the-impact-of-ai-on-the-future-of-insurance
https://www.modernhealthcare.com/article/20180726/NEWS/180729932
http://www.insurancefraud.org/statistics.htm#2
http://techgenix.com/ai-insurance-industry/
https://emerj.com/ai-sector-overviews/artificial-intelligence-in-insurance-trends/
https://www2.deloitte.com/de/de/pages/innovation/contents/artificial-intelligence-
insurance-industry.html
https://www.cbsnews.com/news/digital-disruption-is-rocking-the-insurance-world/
https://www.forbes.com/sites/blakemorgan/2018/05/16/heres-how-iot-will-impact-the-
insurance-claims-process/#6fac85ea366e
https://www.smh.com.au/money/super-and-retirement/a-selfie-could-become-the-new-
way-to-obtain-life-insurance-20170616-gwsl2m.html
14
https://www.the-digital-insurer.com/dia/allianz1-insurance-customisation-in-real-time/
https://www.prnewswire.com/news-releases/clover-health-launches-new-in-home-
primary-care-program-supported-by-genomic-testing-capabilities-300645346.html
https://www.guidewire.com/products/guidewire-cyence-risk-analytics
https://www.mobihealthnews.com/content/beyond-verbal-launches-research-platform-
leverage-health-detecting-voice-analytics-software
https://www.bestliferates.org/blog/2018-life-insurance-statistics-and-facts/
https://www.cmswire.com/digital-marketing/natural-language-generation-accelerates-
content-marketing-but-keep-your-hands-on-the-wheel/
https://www.gartner.com/doc/3611117/magic-quadrant-business-intelligence-analytics
https://www.researchgate.net/publication/323266830_SaferDrive_An_NLG-based_
behaviour_change_support_system_for_drivers
https://www.frugalforless.com/avibra-app-review/
https://www.nanalyze.com/2017/08/nauto-autonomous-driving-data/
https://findbiometrics.com/behavioral-biometrics-smart-bet-for-insurers-
biocatch-512055/
15