Hi
I would like to automate data reporting into text. To do that I would need my Flows to have calculation step in large data sets. For example: What is the sum of January 2026, what is the Variance vs prior month, vs prior year, what is CTG, etc. How can I do that ? Thank you
Hi @dehouckl and welcome to the Quick Community!
From my understanding, you can interact with your data via two approaches: ‘Quick Suite data’ or ‘Dashboard and topics’.
Within the Dashboards and topics step, you can attach either a dashboard or a topic to the step and potentially ask it to perform your calculations within the prompt.
Within the Quick Suite data step, you can attach either a space or a knowledge base to this step and potentially ask it to perform your calculations within the prompt as well.
The difficult part is that you cannot directly attach a large dataset to a flow unless it is done through knowledge spaces, spaces, dashboards, or topics.
Hope this helps!
Thanks Jacob!
My problem is that Quick is bad at making calculations. Is there a way for me to “ask” to code a formula to make the calculations ? The LLM itself can’t handle the calculation.
Thanks
Louis
Hi @dehouckl ,
Can you elaborate on what you mean by asking to code a formula?
You can potentially try to explicitly state the calculation formula in the prompt and that might get you closer to your desired output. Another option you can do is create a dashboard that has all your relevant information as well as your desired calculated values so that way the actual calculations are being performed on the dashboard level and the flows is simply retrieving the value. Doing the calculations on the dashboard level would allow you to have more control over the calculations and ensure that the output is more accurate.
Hi Jacob. I see thank you. The problem with the dashboards is that it would take more time to implement. My goal would be to ask the LLM to process calculations not through “LLM” but through Python for example. Example: What is the average monthly variance of customers x,y,z in 2025 ? That is nearly impossible to do for a LLM in a large data set, but if I am able to ask QuickSuite to extract the data and code the formula instead it would work I believe ? Do you know how could I query this kind of requests ? Many thanks!
Hi @dehouckl ,
Upon research, you can potentially use Quick Automations to write your python.
In Quick Automate, you can use code actions where you manually write your own Python code as part of the automation workflow. So you would build an automation, add a code action step, and write your custom Python formula to create your calculations. From what I have research you would have full control over the logic, the data extraction, and the calculations. Quick Automate also supports data table actions for working with structured datasets, so you can load their data, run your custom Python code against it, and output the results. Here are some helpful resources to check out:
One thing to add is that I have not personally tried this approach so am unsure if this would give you the desired output that you are looking for. Try it out and let me know!
