Several economic crises in the recent past have triggered a regulatory tsunami aimed at improving investor protection and monitoring systemic risk. Banks are having to comply with complex regulatory requirements, often across geographies. A multitude of regulations like MiFID, FATCA, COREP, and FINREP, initiated by different regulatory agencies like FRB, FCA, and FINRA, require financial services organizations to submit reports in a variety of formats such as XBRL, XML, and FTP. This has made navigating the regulatory landscape a tedious exercise, not to mention the dire consequences of non-compliance.
With manual reporting activities presenting challenges around data accuracy and inconsistency, audit trail, timeliness, and operational risk, organizations are looking at technology solutions to ease out the regulatory reporting function. Digital technologies, cognitive automation, and cloud-based solutions have the capability to change the fundamental premise of how reporting is accomplished.
Raise Your Financial Reporting Game with Cognitive Automation
For static reports generated at regular intervals, cognitive automation helps execute repetitive tasks and validate the output to ensure consistency within the purview of the materiality limit and acceptable tolerances. For custom regulatory reports, self-learning and pre-programmed scenarios facilitate report generation with minimal human intervention. This frees business users allowing them to focus on report outcome analysis before statutory submission and initiate remedial measures if need be. Based on the analysis, stakeholders can make informed business decisions and initiate mid-course corrections.
Deploying automated financial reporting systems requires some groundwork. Three key aspects that need to be taken care of to future-proof machine-moderated reporting activities are:
1. The data present in a bank’s IT systems should be complete, accurate, and reliable
2. All possible reporting scenarios must be captured
3. The underlying algorithms should be minimally hard-coded so they are flexible enough to adapt to changing business realities
Data Management – Take a Pause, and Think it Through
In the midst of all the euphoria around leveraging cognitive automation for regulatory reporting, financial services firms must not forget that for an automated reporting system to deliver on its promise, the input data has to be of high quality. Data sufficiency is also an important aspect, and so is the mechanism to tackle errors and deficiencies. Another pertinent question that firms must address is about the heterogeneity of source data. With source feeds being a combination of home-grown, legacy, and state–of-the-art systems, how do you ensure technology compatibility with the end-state cognitive automation framework?
Additionally, financial services firms will have to ensure that the data has been optimally normalized to minimize redundancies, and foster adaptability to changes. Now, this is a mammoth exercise since all geographies, business units, and their respective data feed providers will have to align in this direction.
Radical Changes Call for Elaborate Change Management
With cognitive automation taking over the financial reporting function, the work composition of several business users will undergoes a dramatic shift. For instance, the teams engaged in regulatory reporting may have to apportion their efforts on improving overall quality and accuracy of reports from a business standpoint, as also reducing errors and simplifying reporting templates, and so on. It is therefore imperative to have a buy-in on the approach from all business units to formalize a change management program. Also for business units, unified standard operating procedures (SOPs) would help to have a seamless transition to the new way of regulatory reporting. In the same stead, corroborative changes are expected in current BAU formats of approvals, hierarchies, and workflows.
Do the Wise Thing – Start Small
Considering the huge investment it entails – in terms of technology, personnel, and business time – a cognitive automation initiative naturally involves a lot of cost-benefit analyses and trade-offs assessments at various levels in an organization. And, why not? It’s only wise to ascertain the short-, medium-, long-term benefits of automated reporting as opposed to continuing with business as usual.
In this backdrop, it is best to conduct a proof-of-concept (PoC) using a representative bunch of regulatory reporting templates so as to gauge outcomes, conclude lessons learned, and establish remediation measures. Banks must also look at bundling of reports based on a set of criteria, such that the underlying algorithms can be optimally conceptualized and implemented.
The success of this exercise is crucial, because the ultimate objective of cognitive automation is to build a financial reporting system that is in a self-learning mode, where in a given time horizon the reports become self-driven and smarter. The end system in pursuit of cognitive automation should be able to interpret situations not encountered before, with manual intervention as an exception. This must be based on the previous history of reports, reporting events, and the built-in self-learning by the cognitive automation framework. Over time, this can deliver tangible benefits to business users in terms of costs and reporting accuracy. What do you think? Do share your views in the comments section.