How to save a million dollars on bank operating costs

Verifying suspicious transactions means a tedious information search, collection and analysis. Tedious and expensive at that. But there’s a better way. By Tomasz Czech, head of R&D, AI/ML practice, and Paweł Kryszkiewicz, product manager, financial services, Comarch

15 September 2020

An AML analyst spends an average of an hour analysing a suspicious case. Each time they have to perform a series of simple, repetitive actions. Search for a given client in the database of suspects, verify if the client is not on the warning lists, check what is known about the client from the internet.

In some institutions, even a dozen or so such checks are performed. Many simple operations should be performed in spreadsheets or other forms where the analysis results are collected. Documenting the results of such analysis is necessary for audits. Oftentimes, one has to sort data according to counterparties, summarise turnover on accounts or verify whether the whole transaction has been properly imported.

There is nothing wrong with repeatability, after all, most companies want to follow procedures religiously. Unfortunately, in many places routine and the pursuit of the best results creep in. And these, measured only by rigid metrics, often result in an attempt to exploit the system so as to achieve the goals with as little trouble as possible.

Small numbers – big savings
With the use of an RPA tool (Comarch Robotic Process Automation), the time of analysing suspicious cases can be reduced by up to 12 minutes by automating the information search process. In a medium-sized AML department of 50 analysts, it can save up to one million dollars a year.

The average earnings of an analyst vary between $40,000 and $50,000 per year. To this, you must add the employer's costs, the cost of a manager's work per team, equipment, rent, IT support and many others. The amount rises very quickly to $100,000 a year.

With this assumption, a 12-minute reduction in time every hour with an RPA tool will save $1m in a team of 50 people. The tool automatically collects information about specific companies from the internet, searches government databases, and produces a report that becomes the starting point for the analyst's work. By analysing data from different sources and comparing them with each other, it is possible to identify discrepancies between the data in the blink of an eye. 

Flexibility in the selection of data sources for analysis, and the ability to automate the indicated parts of the process not only bring savings, but also reduces the risk of many errors.

Improvement before redundancy
Anti-money laundering, onboarding, anti-fraud: these are just a few of the many processes in banks that can be improved.
As far as the AML process itself is concerned, its individual stages can also be optimised, such as monitoring transactions and customer activity. In this area, the use of AI-supported RPA for the initial analysis of alerts before they reach the analyst can generate further savings. 

Of course, there are tasks that computers will not be able to perform as effectively and accurately as humans for a long time to come. Such tasks include the analysis of collected data. However, in order for a human being to be able to analyse things effectively, machines should help them quickly and effectively handle repetitive activities.

In today's banking, the problem is often an excess of work resulting from many legal regulations imposing additional obligations, and suboptimal IT processes and tools that support them. Therefore, it is worthwhile to use both the tools that automate these processes and those that use machine learning to solve increasingly complex problems.

Learn more about Comarch anti-money laundering software solution

Become a bobsguide member to access the following

1. Unrestricted access to bobsguide
2. Send a proposal request
3. Insights delivered daily to your inbox
4. Career development