A Developers Insight into machine learning in the financial sector: Opportunities Galore for AI in Banking

The Banking Industry as it stands right now, is faced with insurmountable challenges. While the threat of decentralized cryptocurrencies like bitcoin looms large, traditional routes of income for major financial institutions are are also gradually getting clogged.

While banks have typically been wary of introducing AI related software components into their workspace, the present situation is dire enough even for naysayers to realise that cognitive banks are the future.

The industry leader in this field is IBM Watson with their AI being able to address several pain points in the investment as well as banking sector better than any other machine learning based alternative on the market.

Solving the Regulatory Compliance Problems

So what is so special about IBM Watson? Well put simply, the main cash leaks for banks in the past half a decade has been on regulatory compliance with US Banks spending $80.2 billion on it way back in 2013. And Watson can reduce that number quite drastically. IBM’s acquisition of Promotionary Finance Group, the leading consultancy for banks worldwide for compliance related laws, means that Watson can now be trained by the industry leaders on regulatory compliance for financial institutions.

This is particularly important because as these compliance and regulations keep increasing in number as well as complexity, it becomes near impossible for professionals to keep up with all the nuances.Watson’s cognitive learning capabilities mean that now the power of AI can be harnessed to help banks reduce costs, save time as well as remain compliant with the laws.

Resolving Financial Crimes

While regulatory compliance is a big money drain for banks, another related issue that brings down efficiency of banking analysts a lot are false positives on financial crimes. Using the systems of today that do not have the added benefits of neither AI nor Machine Learning, banks are faced with false positives percentages in in the high 90s when it comes to money monitoring techniques for generating financial crime alerts.

 

This is where Watson can step in and dramatically reduce the number of false positives with their new IBM Financial Crimes Alerts Insight with Watson module. This is a cognitive intelligence module that results in Watson using it’s machine learning techniques in conjunction with its regulatory and compliance unit in the Financial Compliance sector to drastically reduce the number of alerts or red flags generated by modern Anti-Money Laundering programmes.

 

Using an advanced AI like Watson not only reduces risk and increases compliance, it also helps bank perform better in KYC data collection. This ultimately leads to reduced costs as there is less manpower involved in the investigative as well as data collection procedures.

Renovating Customer Marketing for Banks

The advent of fintech companies as well as the changing habits of millennials mean that in the next few years banks as well as other financial institutions will feel a huge pressure in maintaining revenue streams. In fact one report goes as far as to suggest that in the next 5 years millenials, who now control a vast amount of the wealth, will do away with the need of traditional banks. In fact 33% of millenials are ready to switch banks within 90 days which just highlights the dire situation banks are in when it comes to customer retention.

 

As the new generation of customers become ever more tech savvy, banks have seen a surge in digital banking as well as a rise in demand for personalised wealth advisors. With the power of AI and natural language processing, IBM Watson Marketing systems can solve these problems for banks by using machine learning and cognitive intelligence methods to analyse the huge amount of data that is already collected by banks on their customers and then use it to deliver a personalised experience for each one of their customers.

 

IBM Customer Insights is one of the global front runners in this category and can help retain at risk customers, anticipate and provide dedicated product advertising to specific customers depending upon a wide variety of factors including economic demographic, age group and others. All of this is part of Watson’s Core Banking features that makes it one of the most advanced AIs in the financial segment.

Revolutionising Communication Channels for Financial Institutions

As part of the hyperconnected customer service suite, banks have to provide a lot of information that is pertinent and also at a moment’s notice. For a human this becomes increasingly difficult as the data sizes just keep on growing and we have seen that AI has taken over this segment of banking in the form of introducing chatbots that are driven by these machine learning platforms. One such implementation is Eva, a joint venture by Senseforth and HDFC has been revolutionising banking in the Indian markets with its intelligent chatbot capabilities.

 

But there are a few limitations to these early chatbot implementations that one which is built on cognitive intelligence like IBM Watson can overcome. Emotions are a driving factor behind many human decisions and Watson’s Alchemy natural language processing helps in making it a more empathetic chatbot that realises the gravity of each situation and then crafts a suitable response. Such an intelligent chatbot is already in development using the Watson platform-Luvo has been in the testing and demo phase for the Royal Bank of Scotland for a while now.

 

And it’s not just one bank, Deutsche Bank has planned on using IBM Watson to provide an AI powered communication platform for their German speaking client. The Royal Bank of Canada has gone one step further and used RBC Embark which is again based on IBM Watson, to better train their new employees and get them familiarised even before they join work. One of the main reasons why so many banks including giant international banks like the Bank of China have embraced AI powered chatbots is due to the ease with which they can be programmed. In fact creating a Watson chatbot is as easy as knowing Node Js and then using a detailed Github tutorial to prepare your own personalised Banking chatbot using Watson Conversations.

What Sets IBM Watson Apart?

While there are a plethora of options when it comes to neural networks, only one of them- the IBM Watson platform truly stands out head and shoulders beyond any other competitors in the natural language processing arena. But what sets Watson apart from the remaining competing neural networks out there? Mainly two things-the hardware and the software behind the project is at a level that we have never seen before.

 

Hardware: Watson is a feat of parallel processing, the massive computing power is drawn from a total cluster of 2,880 Power7 cores that are uniquely designed for this workload. All these cores are distributed into 90 IBM Power 750 servers each of which holds at its computing center a 3.5Ghz Power7 Octa-Core processor with 4 threads running simultaneously on each core. Coupled with all of this we have 16 terabytes of RAM and it’s no exaggeration to say that the Watson hardware is one of the best when it comes to distributed machine learning setups.

 

Software: On the software front, Watson runs on Intel’s Deep QA software. But analysis is only part of the problem. Watson deals with a huge influx of new information everyday most of which it finds in unstructured data formats. To handle all of this data, Watson is dependant on the Apache UIMA which works in close conjunction to Hadoop, one of the most prevalent softwares in the big data stage.

 

Why are all of these so important? Well the hardware specifications of Watson means it makes 80 TeraFLOPS, a score that is highly commendable even for a supercomputer. And while the Watson project started out as a question and answer providing neural network, in its current stage, equipped with the right sensors, Watson can ‘see’, ‘hear’, ‘read’, ‘talk’, ‘taste’, ‘understand’, ‘reason’, ‘interpret’, ‘learn’ and ‘recommend’.
This is the closest we have come to interpreting the human brain and replicating the cognitive decision making models that we have biologically perfected over centuries in an AI. And this is what in my opinion makes Watson so different from the other neural networks out there. While most distributed computing systems are looking to solve a particular problem that humans face, a complete AI like Watson looks to focus itself on being more human and thus it is uniquely capable of working with humans rather than for humans.

 

From a developer standpoint this means that the platform is uniquely capable of reasoning and handling fringe cases by itself which means that QC for software developed on the Watson platform can be more lenient and developers can use more ingenuity to cover broader subjects knowing that machine learning and cognitive intelligence will take care of the rest.

Cognitive Banking is the Future

The basis of a cognitive bank like the ones powered by IBM Watson is not to replace human interaction and workloads by AI. Rather, the aim always remains to reduce the human workload by augmenting human skills with machine learning approaches. While AI and machine learning are the very customer and profit centric front of the new banking revolution, a cloud based cognitive platform like Watson also provides an overall better mobile and cloud platform with increased cyber security. The future of banks lies in understanding and embracing the new data driven economy with AI-s like IBM Watson doing the heavy lifting for them. The future lies in AI and humans working side by side.


Feel Free to share your thoughts in the comment section below.
Don't forget to follow us on Twitter, like our Facebook Fan Page and Add us to your circles on Google+ to keep you updated with the latest technology news, gadget reviews, launches around the world and much more
Tagged with