• Written by Azhar Aliyar, Director

A finance manager reviewing the company's cash flow statement may wonder what data is required for the AI system to provide accurate answers to their queries. The answer lies in the specific systems and fields that the AI deployment depends on. For instance, a shift supervisor checking the maintenance log may need to know the current budget allocation for equipment repairs, which is linked to the general ledger and purchase orders.

The systems

The general ledger is a critical system that provides the foundation for financial analysis. It contains the company's financial transactions, including income, expenses, assets, and liabilities. The budget system is also essential, as it outlines the planned income and expenses for a specific period. Open commitments and purchase orders are other vital systems, as they represent the company's contractual obligations to pay for goods or services. Additionally, the receivables and terms systems are necessary, as they track the amount of money owed to the company and the payment terms agreed upon with customers.

For example, a category manager comparing two suppliers may need to access the purchase order table to determine the total value of orders placed with each supplier. This information can be used to negotiate better prices or to identify areas where costs can be reduced. The AI system can provide answers to queries about the company's financial performance by analyzing data from these systems.

What a missing system does to an answer

If one of these systems is missing or incomplete, the AI system's answer may be inaccurate or incomplete. For instance, if the general ledger is not up to date, the AI system may not be able to provide a correct answer to a query about the company's current cash balance. Similarly, if the budget system is not integrated with the AI system, the answer to a query about the company's projected income may not take into account the latest budget revisions.

A missing system can also lead to incorrect assumptions or estimates, which can have significant consequences for financial decision-making. Most people assume that the AI system can simply fill in the gaps or make educated guesses, but this is not the case. The AI system relies on accurate and complete data to provide reliable answers, and missing systems can compromise this accuracy.

The definitions that must be agreed

Another critical aspect of deploying an AI system for finance is agreeing on definitions for key terms and concepts. For example, what constitutes a "cost center" or a "revenue stream"? These definitions must be consistent across all systems and departments to ensure that the AI system can provide accurate answers. If different departments or systems use different definitions, the AI system may provide conflicting or confusing answers.

Agreeing on definitions requires collaboration and communication among stakeholders, including finance managers, accountants, and IT personnel. It is essential to establish a common language and set of definitions to ensure that the AI system can provide reliable and consistent answers. This process can be time-consuming, but it is essential for the success of the AI deployment.

Quality problems that only surface later

Quality problems with the data can also surface later, even if the AI system appears to be functioning correctly initially. For example, if the data is not properly formatted or if there are inconsistencies in the data, the AI system may not be able to provide accurate answers. These quality problems can be difficult to detect, especially if they are subtle or intermittent.

For instance, a problem with the data feed from the general ledger may cause the AI system to provide incorrect answers to queries about the company's financial performance. This problem may only become apparent after the AI system has been in use for some time, and it can be challenging to identify and resolve. Regular monitoring and testing of the AI system are essential to detect and address these quality problems.

More information on how to implement an AI system for finance can be found on our finance page. If you have any questions or concerns about deploying an AI system, please do not hesitate to contact us.

When AI is not the answer

While an AI system can provide valuable insights and answers to financial queries, it is not always the best solution. In some cases, a traditional reporting or analysis system may be more suitable, especially if the queries are simple or straightforward. Additionally, if the data is incomplete, inaccurate, or inconsistent, an AI system may not be able to provide reliable answers.

In these cases, it may be better to focus on improving the quality and accuracy of the data before deploying an AI system. This can involve implementing new data collection or reporting systems, or improving the existing ones. By doing so, the organization can ensure that the AI system has the best possible chance of success and can provide accurate and reliable answers to financial queries.