• Written by Muhammed Abdulla NC, CEO

In theory, AI should be able to start showing results in finance from the moment it is connected to an organisation's data. In practice, the first fortnight is often spent on something far more mundane: getting the data into a state where the AI can even read it. This is not just a case of loading the data into the system, but also of standardising the formats, correcting errors, and filling in gaps. Most people get this wrong, assuming that the data is already in a usable state, but it rarely is.

Initial Connection

The first fortnight of an AI project in finance is typically spent on data preparation, as mentioned earlier. This involves a lot of manual work, such as checking for inconsistencies, handling missing values, and transforming the data into a format that the AI can understand. For example, a financial analyst might spend hours poring over spreadsheets to identify and correct errors in the data. This process is essential, but it is not what most people think of when they imagine an AI project.

A shift supervisor might check the maintenance log to ensure that all the necessary data is being collected, while a category manager compares two suppliers to determine which one provides the most accurate data. This attention to detail is crucial, as it lays the foundation for the entire project.

The First Question

The first question worth answering is often one that has been puzzling the finance team for a while. For instance, a question like "what is the average payment period for our customers?" can provide valuable insights into the organisation's cash flow. Answering this question requires the AI to have access to a significant amount of data, including payment records, customer information, and accounting details.

Once the AI has been trained on this data, it can start providing answers to this question, and others like it. This is where the finance team starts to see the value of the AI, as it provides them with actionable information that they can use to make better decisions.

Leading Measures

A leading measure should move within a few weeks of the AI starting to provide answers to the finance team's questions. For example, if the AI is being used to optimise payment processing, a leading measure might be the reduction in payment processing time. As the AI starts to provide more accurate and efficient payment processing, this measure should start to move, indicating that the AI is having a positive impact on the organisation.

This is not always a straightforward process, as there may be other factors at play that affect the leading measure. However, if the AI is working correctly, it should be possible to see a correlation between the AI's output and the movement of the leading measure.

Delays and Complications

So, what makes it take twice as long to see results from an AI project in finance? One major factor is the complexity of the data. If the data is messy, incomplete, or inconsistent, it can take much longer to prepare it for the AI. Additionally, if the finance team is not clear about what questions they want the AI to answer, it can be difficult to get the project off the ground.

For example, if the finance team is unsure about what data to collect, or how to collect it, this can lead to delays and complications. It is essential to have a clear plan and a well-defined set of questions to answer, in order to get the most out of the AI.

Some common pitfalls that can cause delays include inadequate data preparation, poor communication between teams, and unrealistic expectations about what the AI can deliver. By being aware of these potential pitfalls, organisations can take steps to avoid them and ensure a smoother implementation process.

Limitations and Alternatives

This is not a solution for every problem in finance. In some cases, the data may be too limited, or the questions too vague, for an AI to provide useful answers. In these cases, it may be better to use other tools, such as consulting services, to provide guidance and support. Additionally, if the organisation is not willing or able to invest the time and resources necessary to prepare the data and implement the AI, it may not be the right choice.

Ultimately, the decision to use an AI in finance should be based on a careful consideration of the potential benefits and limitations. By understanding what an AI can and cannot do, organisations can make informed decisions about how to use this technology to achieve their goals.