Altus Insurance whitepaper spreads - Flipbook - Page 27
Existing, under-utilised AI capabilities.
While much of the discussion on AI is located 昀椀rmly
in the future, there are existing AI capabilities that
can be utilised to create a marked and immediate
improvement to a customer’s claims experience.
The areas where insurers have delivered the greatest
innovation is in using tools derived from the narrow AI
space. Much of the functionality these tools provide
to insurers are built on increasingly intuitive cloudbased services provided by the likes of Amazon
Web Services, Google Cloud Platform and Microso昀琀
Azure, and are therefore relatively easy to access and
implement.
By becoming specialists in the application of these
more targeted AI components, some of the claims
technology providers identi昀椀ed in the previous
chapter, and others like them, have developed
propositions that can really enhance an insurer’s
approach to a claim, particularly when it comes to
analysis and decision making. The key is to view these
providers (and the technology they supply) with a
collaborative, creative mindset and look at how the
capabilities they provide could be the enablers for a
re-engineered claims model.
The emergence of Generative AI.
While AI is used as a catch-all term, not all
intelligence is created equal. It can be broken
down into the following categories:
• Automation – while still relatively new, these
tools will more o昀琀en than not have a minimal
or no use of AI. For insurers, it usually refers to
the use of robotic process automation (RPA) to
replace a human-managed task with a computer
‘bot’ process. The time and cost bene昀椀ts of
automation are well established as it strips
out manual, repetitive tasks but customers are
unlikely to see signi昀椀cant improvements beyond
quicker response times.
As highlighted above, the key dependency for all
of these solutions, and the application of AI more
broadly, is the quality of the data it uses. This will
include the insurer’s own data but is also likely
to incorporate external sources as part of ‘data
enrichment’ processes. In simple terms, this removes
the need to ask the customer or another party a host
of questions as the information has already been
gathered by the tech from other sources.
While this level of AI has resulted in meaningful
change to some parts of some claims processes, to
move beyond this ‘isolationist’ approach towards
a more global digital strategy, there must be a plan
to implement a robust, secure approach to data
management and governance.
An updated approach to data management will need
to be backwards-looking, to impose new structures
on historical data (upon which any AI tool will be
heavily reliant), whilst at the same time aligning it to
improved processes and systems for e昀昀ective data
capture and processing.
• Narrow AI – this refers to the use of AI-powered
tools in speci昀椀c tasks that form part of a wider
process. This includes the use of natural
language processing, sentiment analysis and
computer vision analysis. As we have seen in
the last two chapters, most of the AI solutions
and initiatives that are being deployed or
marketed to the insurance industry fall into this
category.
• General AI – while not yet a reality, this type
of AI refers to a universal model with humanlike intelligence. Whilst this does not yet exist,
the recent advances linked to ChatGPT, Google
Bard et al indicate a leap forward, and partial
progress from narrow to general AI.
27
