As you can use the Chatbot to describe your discovery / visual using natural language, so you can use the Chatbot to describe your calculations (formulas, lists, or columns) using natural language. All you need to do is type a prompt that describes the calculation you want to create into the Chatbot in business language, and the LLM will build the calculation on your behalf.
You can create calculations in both Formulate and in Discover. In Formulate, you are creating a standalone calculation. In Discover, you are either adding an existing calculation or creating a new calculation as part of your visual creation.
Note: This feature requires specific licensing options
Using Existing Business Logic
If enabled, Pyramid adds vectorization of the metric store - allowing users to easily find and reuse metrics in the store and catalog through NLQ requests. The automated vectorization lets the Gen-AI engines automatically find the right existing calculation in the customer's analytics catalog to answer questions the user may ask with NLQ, without needing to fine tune the LLM and without loading the entire metric repository into the LLM request.
Creating Business Logic
Pyramid will also create new metrics and add them into the metrics store based on questions and requests run through the NLQ AI agents and associated LLMs. There are 2 entry points for this functionality:
Formulate
If the user uses the Chatbot during a Formulate session (specifically for Formulas and Lists), the agent can build the syntax for new business logic formulas or logical lists from natural language requests.
For example:
"Create a formula for gross margin %" will generate a formula
(sales-expenses)/(sales).
"Create a listing of my top 10 customers based on sales" generate a formula
TopCount(Allmembers(customers.fullname),10, measures.sales)
After generating the syntax, the user can save the new formula manually using the standard processes
Discover
If the user uses the Chatbot during a Discover session , the agent can build the syntax for new business logic formulas as part of a broader natural language question, auto-add it to the metrics store and display it in the related Discover result. Unlike the Formulate experience, the user is not shown the syntax, the metric is assumed to be a 'one-off' item and is attached to the report. Users can easily switch the generated metric into a shared item for reuse, governance and sharing purposes.
For example
In a classic sales demo: "Show me sales vs gross margin % by product category" will generate a formula for gross margin % (if it doesn't already exist), add it to the metrics store and then add it to a visual that shows sales and product category as well
(sales-expenses)/(sales)
In a healthcare scenario: "Show me admissions, PMPM and Days/1000 by ICD9" will generate 2 formulas for PMPM and Days/1000 is these items do not already exist in the related metric store for the given model
[measures].[paid amount] / [measures].[members]
[measures].[days]*1000 / [measures].[members]