Introductory Demo | Spreadsheet Powered AI
People & Blogs
Introduction
In this article, we will explore the innovative concept of Spreadsheet Powered AI—a method that integrates spreadsheet models into conversations with Large Language Models (LLMs). We will walk through a demo that illustrates how this technology works using a revenue model represented in a spreadsheet.
Understanding the Spreadsheet Model
The spreadsheet we are using in this demo is a basic revenue model with approximately a dozen assumptions set in a blue box. These assumptions are crucial as they drive monthly calculations that include metrics like the number of visitors, registrations, subscribers, and accumulated revenue figures, including Annual Recurring Revenue (ARR) shown at the end of the spreadsheet, specifically in line 43.
Uploading and Analyzing the Spreadsheet
Once we upload this spreadsheet into our Spreadsheet Powered AI prototype, an intelligent spreadsheet engine analyzes the document and provides a preview to confirm that it is the correct file. The engine generates a text overview that encapsulates the spreadsheet's functionality based on the text labels and context it identifies within the model. It recognizes 13 input assumptions that influence the calculations, such as the advertising budget, cost per click (CPC), and registration conversion rates, enabling users to hover over these values for details about their origins.
The engine also identifies 12 calculated values of interest, which range from the ARR to various time series data points, like monthly recurring revenue, allowing users to track these metrics more effectively.
Engaging with the Spreadsheet via AI
One of the exciting features of this technology is the ability to interact with the spreadsheet through conversation. For instance, if we ask the model to visualize specific data—such as the revenue generated from new versus existing subscribers—the LLM processes this request. It recognizes the relevant time series data and generates a line chart comparing the two revenue sources.
If we want to change that chart into a stacked column format, we can easily make that request, and the LLM re-creates the visualization accordingly, even adding a custom title, such as "Our Journey to 1 Million ARR." As we build out this visualization, we can pin it to the sidebar, effectively creating a dynamic data application.
Modifying Assumptions and Observing Changes
Next, we may wish to delve into the assumptions driving our calculations. For example, if we inquire about the conversion assumptions, the LLM provides relevant figures, such as an 8% registration conversion and a subscription conversion of 2.5%. If we decide to change the registration conversion to 5%, we can witness real-time updates in our visualizations and metrics. The ARR number reflects this change, dropping significantly due to the lower conversion rates.
The technology not only allows for this interactive modeling but also documents the changes made, enabling comprehensive tracking of adjustments and their impacts on outcomes.
Goal Seeking with Spreadsheet Powered AI
As we explore further, we might ask, "What would the advertising budget need to be to achieve our goal of $ 1 million in ARR?" This question represents a natural progression in our interaction, showcasing how intuitively users can navigate the model.
Upon inquiry, the engine applies a "goal seek" function to adjust the advertising budget from its previous setting of $ 8,000 up to approximately $ 31,191—effectively recalibrating the necessary inputs to meet our desired ARR target. The LLM provides explanations of these adjustments, clarifying how the input changes influence the revenue target.
Potential Applications of Spreadsheet Powered AI
The potential applications for this technology are vast, ranging from advanced spreadsheet analysis to engaging chatbot conversations about pricing and financial modeling. Our team is actively exploring various applications of this technology, and we invite interested users to sign up for a demo at spreadsheetpowered.ai to experience this innovative tool firsthand!
Conclusion
Spreadsheet Powered AI represents a significant advancement in how users can interact with financial models, streamlining complex calculations and visualizations through intuitive conversations. We look forward to seeing how users leverage this tool in their projects.
Keywords
- Spreadsheet Powered AI
- Revenue Model
- LLM (Large Language Model)
- Interactive Visualization
- Goal Seek Function
- Conversion Rate
- Annual Recurring Revenue (ARR)
FAQ
What is Spreadsheet Powered AI?
Spreadsheet Powered AI is a technology that integrates traditional spreadsheet models with AI, allowing for conversational interactions and dynamic visualizations.
How does the AI analyze the spreadsheet?
The AI engine processes the uploaded spreadsheet, extracting assumptions and calculated values, providing an overview and enabling users to interact with the model.
Can I visualize data through this AI?
Yes! Users can request various visualizations, such as line or stacked column charts, to represent specific data points from the spreadsheet.
How are changes tracked in the model?
Any changes made through the AI are logged, allowing users to see how assumptions impact the overall outcomes and providing a historical record of modifications.
Can I change input assumptions?
Absolutely! Users can request alterations to input assumptions, and immediate recalculations will reflect those changes in real-time.