Interactive Questions and Answers with Pyramid Analytics’ LLM-Driven Chatbot

Chatbots are increasingly transforming the way users interact with technology. In Business Intelligence (BI), they empower non-technical users to generate insights and visualizations effortlessly, removing the need for complex interfaces or manual queries.

A key challenge for BI chatbots is adapting to the unique and varied datasets of different businesses. To address this, Pyramid Analytics developed an NLQ Chat Bot that integrates with any supported database, allowing users to interact with data through plain English questions.

With features like dynamic filtering, sorting, and contextual suggestions, the chatbot makes data exploration accessible to all. It supports refining visualizations and queries interactively, leveraging advanced AI techniques to understand natural language and deliver accurate results.

Hands-On Example: Exploring Sales Trends?

In this hands-on example, we explore the capabilities of Pyramid Analytics’ NLQ Chat Bot using a simple scenario with dummy sales data for the last 12 months. To begin, we prepared the data by creating an Excel file containing the dummy sales figures and integrated it into Pyramid Analytics’ data model.

Next, we created a Discovery in Pyramid Analytics, an interactive space for exploring data, building visualizations, and uncovering insights. With the data ready, we initiated the session by asking the chatbot,“What was the trend of sales revenue, and which month showed the highest growth”. The chatbot responded with a visualization that highlighted the trend and identified the month with the highest growth. A screenshot of the results is shown below:

While the initial visualization provided some insight, we decided that a line chart would better represent the sales trend over time. Line charts are ideal for showing fluctuations and the overall pattern of revenue. We prompted the chatbot with, “Can you change the current visualization to a line chart that displays the trend of sales revenue, with months on the X-axis and sales revenue on the Y-axis?” The chatbot then adjusted the visualization accordingly, providing a clearer representation of the data. Below is a screenshot of the updated chart:

To gain deeper insights into the sales trend, we asked the chatbot, “Can you add trendlines?” Adding trendlines helps identify overarching patterns and provides a clearer understanding of whether sales are consistently increasing, decreasing, or fluctuating over time. This feature is especially valuable for forecasting and strategic decision-making. The chatbot seamlessly added the trendlines to the visualization, making the data even more insightful and actionable. A screenshot of the updated chart with trendlines is shown below:

Now we can easily answer the question: What was the trend of sales revenue, and which month showed the highest growth? The sales revenue has shown a consistent increase over the year, with December standing out as the month with the highest growth, both in percentage and absolute revenue increase.

We then asked the chatbot, “Can you add a 6-month forecast?” Forecasting allows us to anticipate future trends based on historical data, providing valuable insights for planning and decision-making. By incorporating the forecast, we could visualize the projected sales for the next six months, helping to identify potential opportunities or challenges ahead. The chatbot quickly generated the forecast and updated the visualization accordingly. A screenshot of the chart with the 6-month forecast is shown below:

Additionally, using the forecast data, we can address the question: What are the projected sales figures for the next six months? The chart reveals that sales are projected to reach approximately $315.1K by June 2025. The forecast provides a detailed outlook for the coming months, offering specific revenue projections and insights into potential growth trends. Below is a screenshot of the full Discovery:

Pyramid Analytics’ LLM-Driven Chat Bot offers an intuitive and flexible way for users to explore data and generate insights without requiring deep technical knowledge. Even for users unfamiliar with Pyramid Analytics, the chatbot’s natural language capabilities make it accessible.

While the chatbot may not always interpret prompts perfectly on the first try, with a bit of iteration, it consistently delivers the insights users are seeking. This ability to refine questions and visualizations over time ensures that end users can achieve their goals efficiently, making the chatbot a valuable tool for anyone working with BI data.

Ready to unlock the power of data and explore insights effortlessly? Contact us today and discover how Pyramid Analytics’NLQ Chat Bot can transform the way you interact with your business intelligence data. Whether you're a seasoned analyst or just getting started, our AI-driven chatbot makes data exploration easy, intuitive, and accessible - no technical expertise required!