Skip to main content

​
Getting Started

Once you’re in aidnn, you’re ready to run your first analysis in what we call a Notebook. You can launch a notebook from the Overview page. Also, to get back to the top level menu and the Overview page from any place in the app, click on the Isotopes AI icon on the top left corner as shown:

​
Context

You’ll compute Net Revenue Retention (NRR) for a typical SaaS company using datasets from Jan 2024 to July 2025:
  • Download the PDF with product and pricing details
  • Download the XLSX file containing customers, contracts, subscriptions, billings, and credit reversals
    TBD: add the links for these files.
The analysis includes understanding and preparing the data:
  • Parse the PDF to extract product information and pricing (noting which products bill annually vs. monthly).
  • Parse the XLSX file to understand each tab and relationships such as customer_id and contract_id.
  • Use this understanding to design the approach for generating MRR and NRR.
Then, aidnn performs the full analysis:
  • Execute complex joins across all datasets.
  • Allocate revenue from each customer-product combination across future months according to standard accounting rules.
    • aidnn applies a distribution consistent with widely used ASC 606 SaaS revenue-recognition principles.
  • Aggregate data to produce monthly MRR.
  • Break down each month’s revenue into retained, churned, and new revenue by customer and product.
  • Re-aggregate those results to compute total retained, churned, and new revenue for each month.
  • Finally, generate the monthly NRR metric.
aidnn handles all of this automatically, with full transparency and auditability throughout your workflow.

​
Execution

​
Kick-Off The Analysis

All that you have to do to execute the above complex analysis is submit this prompt and upload the downloaded files to aidnn’s prompt window that will launch a Notebook for analysis as follows:
  • Copy and paste this prompt in the chat/prompt window as shown:
    “Generate monthly Net Revenue Retention (NRR) for Jan 2024 to July 2025.     Find the billings data from Snowflake for calculating revenue. Product and pricing data is available in the attached PDF. Calculate MRR values for individual customers and from that calculate the NRR values. Plot NRR as a line chart.”
  • Then click on the file upload button and upload the earlier downloaded PDF and XLSX files as input files.

​
Review the Plan

aidnn will generate a plan in about a min that outlines the detailed step by step process for analysis like this:\ Please note that the actual plan may slightly differ from what is shown above. Please read the plan carefully at least the first few times and make sure that you agree with the approach before approving the plan.

​
If Required, Revise The Plan

Most likely you will not have to do this for this example. However, in case you want to make any changes to the plan, you have 3 ways of doing this:
  1. To change a given line of the outlined plan or add a new step, hover your mouse on each section of the plan to see a pencil button and click on it to edit any step of that section as shown below:
  2. Then click as shown below to add or edit a step:\
  3. Alternatively, you can provide a feedback overall about the plan by clicking on the Feedback button as shown below:
  4. Upon making changes to the plan, the Approve button above will change to Replan (an arrow if you click on feedback) and you need to click on it to generate the new plan.
  5. Finally, once you are happy with the plan, click on Approve to submit the plan for execution.

​
Plan Execution

Once you approve the plan, a step by step execution of the plan happens as shown below: You can click on each execution tab to view the following:
  1. The input data extracted from the input files uploaded
  2. The processing done at each execution step including the details of any mathematical formula or other transformation applied. One of them is shown above.
  3. The summary of the execution step e.g. any lines filtered, the number of input rows found, the number of output rows produced etc.
  4. Most importantly, it shows all the intermediate data produced for you to view, validate and audit in future.

​
Results Page

Once the execution completes successfully, a result page summarizing the results is generated. It shows the line chart for sure. However, it also shows key samples of the underlying data as well as an overall narrative about the analysis including any anomalies discovered, any insights found or any recommendations for further analysis.

​
Structure of An Analysis

Now, the analysis that you just completed is the first analysis in what we call a Notebook. An analysis has 2 parts:
  1. The analysis Quad(s) on the left side inside the Notebook that contains your prompt, the thought process, the plan, the execution steps and the results generated.
  2. The analysis Details which contains the details of each of the above parts of the quad. For example, it shows the detailed plans and enables you to edit the plan. It shows the details of the execution steps for you to access any intermediate or final results.

​
Continued Further Analysis

You don’t need to stop right here. You can continue to do further analysis. For example, you can ask it to plot the underlying MRR values monthly.
In order to do this, you need to first click on the “Results” section of the analysis quad on the left panel. Then scroll to the end of the analysis details page. You will see a prompt window available here where you can submit a follow up prompt and do your next analysis. All your prompts and the associated analysis will be contained inside the current notebook. You can access this notebook at any point of time from the Notebooks tab available on the Overview page’s left side menu.