In the world of finance, reconciliation is the process of comparing two related sets of records or two accounts at the end of a specific accounting period to find out if account balances are matching in both records. Periodic reconciliation of accounts is an important accounting process and is mandated by regulators to identify accounting discrepancies before they lead to serious book-keeping issues. Timely checking of accounts will identify data issues like duplication, omission, date error and amount error. In many use cases, account reconciliation will spot signs of fraud or manipulations (unauthorized transfers, unauthorized withdrawals, etc.) early. Any discrepancies in reconciliation should be investigated and, if needed, adjustments are to be added to balance both accounts.
The complexity of reconciliation varies with the type of accounts, nature of business and the volume of data to reconcile. There are different reconciliation types of varying complexities such as bank reconciliation, credit card reconciliation, positions reconciliation, custodial account reconciliation, point-of-sale (POS) reconciliation, balance sheet reconciliation, suspense account reconciliation, etc. Depending on the complexity, different businesses use different methods to reconcile accounts from paper ledgers to reconciliation software. With increasing digitalization, paper records are not seen anywhere today. Businesses of all scales are using tools such as spreadsheets, accounting software and robotic process automation (RPA) to address reconciliations of various complexities. Having proven its capabilities across industries, modern technologies such as AI and machine learning are also being used widely, especially for doing complex reconciliations, which would otherwise require a lot more time and huge resources. Here, we are trying to explore various tools used by organizations to reconcile accounts and detail the advantages and disadvantages of each.
Spreadsheets: Outdated Math of Rows and Columns
Spreadsheets have a wide range of use cases including reconciliation. These work as good tools for data analysis and recurring calculations, especially for small organizations. According to surveys, a large portion of financial institutions are still using MS Excel as their reconciliation tool. However, they may prove unsustainable in the long term given the ever-increasing breadth and depth of data reconciliations required. Given below are some of the disadvantages of using spreadsheets as an end-to-end reconciliation tool.
- Manual Workflows: In most spreadsheet reconciliation cases, input data from various sources is entered manually. This is a cumbersome, time-consuming job and is prone to numerous errors.
- Lack of Scalability: With spreadsheets, it is very difficult to handle the ever-growing size of data.
- Loss of data: Spreadsheet owners and users may change, leading to the loss of business-critical information.
- Failure of Formula: Formula written for reconciliation records may fail to work or process when data volume crosses certain limit.
- Lack of Adaptability: System upgrades, key user resignation, changes in regulations may lead to a complete overhaul of the reconciliation process, and spreadsheets may not be able to handle the same.
- Lack of Auditability: Document version history may not always help for auditing purposes and key document changes may go untracked.
- Lack of Analytics: Daily or month snapshots of records produced with spreadsheets may not help reconcilers in modern-day reporting responsibilities. Business currently needs in-depth reconciliation analytics.
Rules-based Solutions: Automation not in Full Scale
Packaged as part of accounting software suite or as a standalone reconciliation tool, there are many software solutions positioned to do large-scale reconciliations. They were able to address some of the issues and drawbacks of spreadsheet-based reconciliation such as manual entries, lack of scalability and lack of analytics. Integrated with ERP solutions in most cases, these tools could also automate a large number of reconciliation workflows with pre-set rules, thereby reducing errors caused by manual entries. With ETL processes, they could better reconcile a large amount of data in a fast manner, increasing the productivity of the reconciliation staff. They also offer enhanced analytics, better user controls and enriched dashboards. However, with reconciliation processes becoming more and more complex today due to disparate data sources and new regulations, their sustainability raises questions. The major drawbacks of these software solutions are:
- Disparate data sources: Adding new data sources may require a large amount of reengineering work. New regulatory standards such as Basel III and MiFID II have significantly changed the scope of reconciliation, mandating financial institutions to reconcile data stretching to more than 65 fields.
- Increasing matching complexity: Rules-based record matching may not always work with new asset types (in financial services) and deals involving complicated calculations.
- Troublesome Exceptions Handling: While RPA solutions could handle matching, exceptions/breaks management is still laborious and costly. Many organizations are finding it difficult to resolve breaks on time and meet compliance standards.
AI-based Reconciliation: Intelligent Automation in Play
As in the case of any other processes, AI and machine learning are revolutionizing the way businesses reconcile data. Modern solutions help financial services maximize efficiency and effectiveness in the reconciliation process. The next-generation software solution adds value in the following ways.
- Fully Automated Matching: It completely automates the matching process in a reliable, scalable and sustainable manner beyond the scope of existing reconciliation solutions. Our reconciliation software can take up large volumes of data additions, do the necessary clean-up and enrichment to produce better matching results. RS can connect to almost all data sources, including existing structured data sources and ETL layers. Our matching engine uses machine learning and distributed systems to rifle through billions of transactions and enhances the match performance and accuracy, without manual rules or updates. It also handles complex match cases (one-to-one, one-to-many, many-to-many), and overcomes matching inefficiencies.
- Automated Exceptions Management: Exceptions handling is still a manual process in the industry. Breaks are investigated and reconciled with the help of numerous staff. Tookitaki RS utilizes machine learning models to understand and learn from previous reconciliation cases, and then predicts both known and unknown exception cases without human intervention. It successfully resolves the exceptions by improving break classification according to a client’s business requirements. It can automatically build exceptions models from historical data generated by the rules-based system.
- Audit Trail: Audit readiness is important for any reconciliation solution today. Data changes, reasoning for matches or breaks and steps taken to resolve a break should be visible to auditors. Our RS solution provides a detailed audit trail as our patent-pending explainability framework helps users understand the rationale behind a break or match case and address it accordingly.
- Seamless Integration to Existing Systems: Tookitaki RS can seamlessly be integrated to existing ERPs, internal systems and platforms.
- Automated Model Management: Our automated model management framework creates machine learning models easily. Its continuous learning ability keeps the performance of matching and exception handling engine intact with new data types and increase in data volumes. Tookitaki RS also features a strong analytics layer that brings detailed insights to drive business outcomes.
- Wider Application: Built to handle almost any reconciliation type, Tookitaki RS covers sectors such as capital markets, banks and corporates and can support asset classes like equities, forex, cash, commodities, and derivatives.
Talk to An Expert!