SocGen Goes Next-Gen with Tookitaki Reconciliation Suite

Société Générale leveraged Tookitaki’s Reconciliation Suite to automate the bank’s existing break reconciliation system

Key Challenges Addressed

– The bank’s finance department was unable to handle new breaks unless code is updated
– Current system cannot detect underlying data changes
– Breaks for new back-offices needed explicit coding
– Existing system was extremely resource intensive

Business Benefits Achieved

– Automated handling of break resolution
– Accelerated RCG process
– Significantly minimized manual effort
– Reduced operational risk
– Established an audit trail

Société Générale’s short-term lending and borrowing business unit wanted to improve its reconciliation process, leveraging the possibilities in machine learning. The Department of Finance (DFIN) in the bank used RCG, which is the process of doing reconciliation between the accounting and inventory data to find out discrepancies which would have been caused due to manual or system errors. Exception handling in the system was managed by writing rules to identify the cause of the break. The bank wanted to largely automate the break reconciliation process with the help of Tookitaki’s Reconciliation Suite.

The Roadblocks Ahead

The bank’s system was unable to handle new breaks until code was updated manually. In addition, explicit coding was required for new back-offices to handle exceptions. Another problem with the process was it could not detect underlying data changes, which affected the quality of the break reconciliation process. In order to rectify new breaks, analysts had to update code very frequently. This was always a time-consuming and resource-intensive process.

Finding the Solution

After careful consideration of a number of solutions, Société Générale opted to proceed with Tookitaki’s Reconciliation Suite powered by machine learning. The solution’s feature set includes the ability to create accurate models 100x faster, a robust model audit feature and self-learning capabilities to ensure models are accurate and updated. The project went technical live in September 2017. The overall model accuracy stands at 98% in production.

How It’s a Job Well Done

Tookitaki provides a faster time-to-market as there is no code development required. With the implementation of Tookitaki Reconciliation Suite, the bank’s break resolution system has been updated with richer and more comprehensive rules. The new system, which can automatically evolve models based on underlying data pattern changes, has achieved high accuracy levels over time. In addition, Tookitaki provided full explainability of decision criteria for audit purposes, making the auditor’s job easier. Further, Tookitaki significantly improved the quality of the investigation process while substantially reducing investigation time.

Tookitaki-Société Générale Collaboration

Société Générale’s engagement with Tookitaki dates back to May 2016, when the bank launched its Catalyst programme. Tookitaki is one of the few start-ups, with whom the bank collaborated to automate break reconciliation process in finance. The program aimed at co-creating solutions to meet customer needs by collaborating with disruptive technology startups. The program provided startups with an apt platform to test, iterate and scale their products or solutions with a real scenario, with guidance, coaching and mentoring from Societe Generale’s experts.

Categories: Case Studies