AI training in finance and controlling

First comparable data

Before the session, we select the type of report and determine which data can be fed into the tool. The exercise starts with describing the period, unit, and table range. Participants check the calculation results in the worksheet or another agreed source. AI can help formulate questions, but the smoothness of the comment does not guarantee a correct calculation.

Let's assume that one file contains the costs of the entire quarter, and another only the last month. The generated comment describes the difference as a drop in costs. The participant should pause the interpretation and point out the needed comparative material. The correction also includes eliminating the cause of the drop, which the data do not confirm.

Comment for the person responsible for the budget

The second part may involve converting the analyst's note into a shorter explanation. The author maintains the information about the gaps and separates the observed change from the hypothesis about its cause. The author prepares a question for the cost owner, when the table is not sufficient to resolve the issue.

The facilitator refers to specific comments: incomparable periods, omitted unit, or overly certain conclusion. After the correction, the participant receives new material, where the periods are consistent, but part of the values are given in thousands. The independent task checks whether the participant notices the change in the unit and considers it in the comment.

If the group is preparing materials for the management, you can practice a summary with a reference to the source table. The scope of the training does not include investment decisions or tax opinions. It also does not replace the agreement of data in the company's financial system.

Group selection

The controller analyst and the person presenting the results may work on the same material, but they need different exercises. Before the evaluation, we establish roles, used tools, and experience with AI. Beginners get time to familiarize themselves with the environment; regular users can work on a more complex comment.

To prepare the program, it is enough to describe the report. The sessions themselves can use fictional data, without information about customers or employees. The company approves the scope of materials and the tool allowed for their processing.

Evaluation and further work

In the offer, we define the length of the program, preparation of materials, and the way of discussing the work. Additional sessions and support after the workshop require agreement. Licenses are purchased by the client, and configuration of the environment or data integration is a separate order.

Describe the report, its recipient, and the most common questions for the comment. For teams working with market information, the addition is a guide for AI training for the treasury department.

How much does the training cost and when does it pay off

The training starts at €1,150 net per day. We calculate the return on one process, without assumptions about the entire company.

According to a study published in Science (2023), the time for writing tasks with ChatGPT has dropped by 40%. That is why the scenario adopts a cautious weekly hour and checks it with measurement.