How AIOps is Modernising the Financial Sector

By Kunal Agarwal, CEO & co-founder, Unravel Data
Kunal Agarwal, CEO & co-founder, Unravel Data, comments on the effects of DevOps and the move to cloud platforms on the financial sector As the mat...

Kunal Agarwal, CEO & co-founder, Unravel Data, comments on the effects of DevOps and the move to cloud platforms on the financial sector

As the maturity of cloud platforms has developed an increasing number of sectors have seen the need for migrating their data workloads to the cloud. While the financial sector has traditionally been slow to move their on-premise, legacy workloads to the cloud, many major banks are now finding that cloud migration has direct relevance to their business use cases.

Two of the main areas where banks are beginning to use big data are in fraud detection and compliance application performance failures. For these specific use cases the cloud already has significant advantages over an on-premise environment. However, the increasing adoption of AIOps has compounded these advantages to such an extent that financial institutions leaving their workloads on-premise will soon find themselves outmoded. 

Before looking into how AIOps is restructuring financial data applications however, let us first look at the inherent issues of running them on-premise. The first is speed; typical on-premise applications in finance are slow and subject to frequent crashes. The key issue is that remediating these issues is an intensely time-consuming task for data teams as they need to manually sort through copious data logs to find problematic issues. This process is usually on the time-scale of weeks, and unfortunately results in application down-time. Even once the problem has been addressed, trial-and-error processes to ensure the issue has been resolved add several more weeks until normal operations can be resumed. Another consideration is how difficult it is to monitor issues at the cluster usage level. Frequently, data teams will have little visibility over how computer resources are being used and, as a result, optimising data applications is a constant challenge. Due to this lack of visibility, compute utilisation issues can only be identified when a critical data application fails.

With such a comprehensive host of issues on-premise, the incentive for financial institutions to migrate to the cloud is clear. The primary advantage of hosting these applications in the cloud is the enhanced visibility it provides when well managed. Whether these applications are using Hive, Spark, Workflows, Kafka, etc. the ability to monitor their performance in real time provides data teams with invaluable insights that can be used to create reliable performance. For specific use-cases like fraud detection where the inputs of streaming data are so large, the cloud is even more integral in ensuring teams still have visibility over applications and detailed insights gained. With these insights, applications can be optimised to boost performance and reduce wasteful resource consumption.

This is where AI and automation are a boon to cloud workloads. AIOps promises to enhance or replace a variety of IT operations processes through combining big data with AI or machine learning. For the financial sector specifically, this holds obvious appeal as expensive, challenging, and time-consuming problems in big data deployments can be addressed. With AI capable of independently diagnosing several root cause issues, the burden for data teams is reduced as there is no longer the need to manually sort through data logs. Moreover, with automation also being able to provide notifications for specific failures, there is the option of providing automated fixes in these specific instances.


For financial institutions looking to enjoy these benefits, however, the journey to the cloud can seem confusing. As such, the biggest obstacle is often the planning phase. In this process, organisations need to map out which applications are suitable for the cloud and which should remain on premise. Another objective of this phase is to anticipate how these applications should be configured based on specific instance type recommendation. What these early decisions allow for is accurate cost and consumption forecasting. Importantly, by providing critical, data-driven insights instead of relying on guesswork, the decision making timeline can be accelerated and the cloud migration and realised sooner.

To conclude, undeniably AIOps will be essential for financial institutions moving forward. As such, organisations should look to begin planning their cloud migration early to realise the advantages that AIOps has to offer in their deployments. This will broadly reduce their costs, increase efficiency and allow them to redirect resources to value-on initiatives instead of fire-fighting outmoded on-prem workloads.

For more information on all topics for FinTech, please take a look at the latest edition of FinTech magazine.

Follow us on LinkedIn and Twitter.


Featured Articles

WE’RE LIVE! FinTech LIVE Dubai

Back for another day, this time in Dubai! FinTech LIVE Dubai is LIVE, don’t miss out on your chance to hear from Swift, HSBC, Mastercard and many more

Amberdata: RWA tokenisation gains significant momentum

Explore the world of RWA Tokenisation and why finance professionals are investing in the technology for sustainable growth and risk mitigation

WE’RE LIVE! FinTech LIVE Singapore

Kickstarting the year of events for FinTech LIVE, FinTech LIVE Singapore returns featuring speakers from Amberdata, Standard Chartered, ING, WeLab and more

FinTech LIVE Singapore: Just One More Day to Go!

Financial Services (FinServ)

Top 100 Women 2024: Allison Paine Landers, UBS - No. 10


Top 100 Women 2024: Akila Raman-Vaseghi, Goldman Sachs No. 9

Financial Services (FinServ)