Empower IBM TM1/Planning Analytics Users with AI Generated Documentation

Recently, a user asked me for a detailed explanation of how a particular TM1 rule worked for one of our General Ledger accounts. He had two questions:

1. how is this calculated?
2. where else is this being referenced within the TM1 Rule?

For his first question, I told him to use the Trace Cell feature to help him understand the TM1 rule calculation, but since he's not a TM1 developer, he understandably found the content difficult to grasp:

Answering the second question required a simple text search for the Account within the TM1 Rule to identify where it was referenced, but this process was tedious and time-consuming.

I set out to provide users with easy access to the answers to these two key questions. By leveraging the TM1 REST API and optimizing prompts for the ChatGPT API, I automated a process generating this information and storing it as text in a dedicated cube. This data is then easily accessible through a TM1 Drill-through process, streamlining the user experience. Additionally, other AI APIs can be used to achieve the same results, offering flexibility in how this solution can be implemented.

Here are examples of clear and user-friendly explanations the Drill-through process will display.

• The Rule calculation for the Attrition Expense account answers the first question: how is this calculated? It details the specific expense accounts involved, along with the Headcount Assumption used, providing a comprehensive view of how attrition-related costs are derived.

• The Salaries & Wages account highlights its role as a key component in the Attrition Expense calculation. It answers the second question: where else is this being referenced within the TM1 Rule, providing valuable insight into its broader impact on related calculations.

• An account may have different rules depending on the business context. The COGS account illustrates how its calculation varies between department DA30 and DB30, ensuring that department-specific differences are accurately captured and understood.

• This process will store both answers to "how is this calculated?" and "where else is this being referenced within the TM1 Rule?" in the same Drill through area. Here is an example of how Bonus Monthly Max is calculated and where it is referenced in other calculations.

Finally, nearly any language can be implemented:

If any of your Rules change, simply run the process to automatically update the explanations. Once everything was set up, running the process became effortless. Each cube took about 5 minutes to execute and store the results. If you're interested in learning how to set this up, feel free to reach out—I’d be happy to help you get started!