Celebrate Your Worth

Is the Financial Services Industry up for the Challenge?

Automation is great – as long as the immediate obstacles are tackled head-on.

Among the wide range of emerging technologies, artificial intelligence and digital labour seems to be the most eye-catching going ahead. Their potential is unmatched – but to be fair, so are the ensuing challenges. Recent advancements, including robotic process automation and intelligent process automation (RPA and IPA) are slated to solve major business problems in the financial services industry. These solutions, however, may often come with several real-world problems of their own – such as governance and control – which need to be addressed immediately if they are to become mainstays for the future.

Decisive action will be mandatory
Most of today’s automated systems require humans to train them. However, this could change soon. With emerging technologies such as intelligent process automation (IPA), including natural language processing, auto process discovery and machine learning coming into the foray, these advanced tools will not only learn from prior decisions and data patterns, but also learn using artificial intelligence to handle much more complex assignments.

‘Augmented intelligence’ will be the next step, offering tools to assist humans make decisions – whilst machines will learn from these interactions as well. Organisations can also consider AI as a way to develop predictive analytics and customize product design to improve outcomes – such as reduced accident rates. For these technologies to succeed in the long run, however, three major problems demand urgent attention.

● Firstly, with many financial institutions eager to move towards digital labour, human workers are starting to feel threatened. This has to be sorted out quick and calls for focus on ‘people issues’. Transparency will be key and will facilitate employees to understand how exactly their jobs will change in the coming years and allow them to seamless adaption to the new digital-normal.

● Secondly, there needs to be a new and independent IT Audit team. Although this requirement isn’t glaringly obvious at the developmental stage or in a testing lab, yet it becomes imperative once one reaches production levels. An AI audit team – independent from its creators and implementers – will be focussing on controls. Again, transparency and accountability will be key. Answering questions like how and why certain algorithms are reaching certain decisions, will be crucial in guaranteeing AI longevity.

● And finally, data management.
With great Data comes great Challenges
Data will be the driver of industries and data-driven decisions and insights will be the way ahead. In fact, the rate of data creation has now become so very high, that we could see an annual 44 zettabytes of data being created just this year (a petabyte is 1012 bytes and a zettabyte is a billion petabytes, or 1021 bytes). This is, of course, an inordinately large amount of data, and firms need to focus on finding the best possible ways of storing, classifying and harnessing this massive resource.

It is, however, alarming to note that only 37% of major financial services’ professionals today state that they would be making their next big decision using internal data and analytics. This is especially alarming in light of a recent PwC report that reveals financial institutions – both retail and commercial – have more data on their customers than anyone else.

However, utilising all this data meaningfully is still a far-fetched reality, with most estimates stating that businesses appropriately use only 0.5% of the total data available to them. A greater portion of the unused data gets stored in silos or in non-compatible formats for their entire life-cycle. This is a costly affair for most firms, as storing this data with adequate privacy settings is both expensive and time-consuming.

Converting data into insights, however, needs to be the real goal – and that becomes even more difficult because almost all available data is unstructured. In order to prepare data for machine learning, they need to be properly sourced, labelled and organised. This includes curating unstructured data such as audio, images and videos as well. Focus needs to be on procuring and using ‘lean data’ whilst maximising value (returns) and minimising waste (efficiency).

This is going to be the primary challenge of this coming decade. By combining business domain, analytics and artificial intelligence, financial institutions will now look towards harnessing data and generating business value. This will, of course, require its own team of AI experts who understand algorithms and can develop new techniques, as well as data scientists/engineers who are adept at using machine learning systems and Cloud technology. The next few years will be spent by financial institutions with focus on recruiting, training and growing teams with these profiles. Make no mistake, automation is great – but only as long as the immediate obstacles are tackled head-on.

Leave a Reply

Your email address will not be published. Required fields are marked *

*