Page 131 - Building Big Data Applications

Chapter 7 Banking industry applications and usage 129

AI and big data for banking applications autonomously spot patterns humans cannot
see. It can also self-correct to keep on improving and adapting to changes in the market.
Which brings us to the term that comes up again and again in machine learning and AI
personalization.
Personalized drug treatments, ad campaigns, and now an AI system that can help a
bank create and recommend better banking products to customers on a personalized
basis. Convincing a customer to plan or provide them with decisions already made for
them are both tricky and will need to learn over time. AI and machine learning can help
marketers to target high potential customers at a time when they are most likely to
respond favorably. However, one application that will no doubt excite a lot of interest
among banks and customers is fraud detection. Of course, an unwary individual could
lose a few thousand dollars by being too relaxed with their bank details, but an insti-
tution, with an increasing amount of valuable company data being stored online, could
be a victim on a much grander scale. And this is where machines really do learn: by
recognizing established threats or potential threats and adjusting to new ones.
Technology implementation within applications today uses AI techniques to analyze
internal, publicly available, and transactional data within a customer’s wider network to
spot rogue behavior has been piloted. Cybersecurity threats have gotten more complex,
clever, and quick, and machine learning that can adapt will be invaluable, where we see
applications including Apache Metron and Cisco Umbrella. Facial recognition, voice
recognition, or other biometric data can be added to market movements and trans-
actions to develop a data mountain accessible and understandable by machines that can
reveal patterns that may be threatening. These data sources, of course, have applications
not only to security but to customer relations and marketing. These are all trends of what
is happening within banking today.
The multi-billion-dollar parts of finance: loans, insurance and underwriting are highly
competitive businesses, and the information to be processed and managed to make
these segments effective is complex and challenging. This is another interesting appli-
cation where machine learning and neural networks have transformed the processing,
increased the availability of data, and associated process reduced time to process and
have improved the data lineage and overall process.
Age, health, and life expectancy are a changing constant and they have all imposed
requirements that need to be managed and adjusted along the way, but how about looking
at the habits of a certain age group in a certain area over a long period to assess otherwise
unnoticed risks or benefits? Do they drink too much in one county? Do they live in a city
with access to medical marijuana? Do they spend more time on computers or mobile
devices online? Have healthier lifestyles? Manage their money better than a similar group
in another town? How precisely can these elements be assessed, calibrated, and used? And
can they be scaled up to millions of examples of consumer data and then applied accu-
rately to insurance risks over an entire population? These are all questions that have been
a tiger’s tail chase for banks. Today with the emergence of algorithms, neural networks,
and natural language processing (NLP) applications, in the backend infrastructure we
constantly attempt to understand natural human communication, either written or
spoken, and communicate in return with humans, using similar, natural language. AI and