Generative AI with Planning Analytics and WatsonX

Generative AI itself is a type of artificial intelligence that can create new content like text, images, audio and even code (check out WatsonX code Assistant). With that intro, I am getting straight into integrating PA/TM1 with WatsonX.AI for text generation based on the cube (financial) data.

What is watsonx.ai?

We cannot talk about generative AI without ChatGPT. Think of watsonx.ai as an ENTERPRISE grade generative AI studio. I will write a separate blog on the features of watsonx.ai but for now remember that we are going to invoke a Large Language Model (LLM) in watsonx.ai to generate text and the prompt (input) will be cube data.

Step 1 – Create Project in the watsonx.ai. Use your IBMCloud login to open watsonx.

In the project assign Watson Machine Learning service – for this example I am using a lite version which is free of charge. We can assign it in the Service & Integrations section of the Manage tab.

Toggle to the General section in the above section and get the Project ID which we will utilise in our TM1py Rest API call.

Step 2:? Generate your IAM API Key from your IBM Cloud login. This will be used to authenticate for login.

Step 3: Create a python file. Few snips and explanations. I am using llama-3-70b-instruct LLM for text generation.

Then have used a Cube view as input prompt to generate text which is shown on the right-hand side. The button invokes a TI process which in turn executes python script which uses TM1Py integration with watsonx.ai

There are certain use case which can be achieved in the office of Finance with Planning Analytics and Large/Small Language Models. More to follow in the upcoming blogs. If you are interested in PA/TM1 GenAI please reach out to your IBM representative, alternatively you can DM me.