The next generation of Business Intelligence will be conversational.

Imagine chatting with your data the same way you chat with friends on WhatsApp. In this blog post, we’ll explore the revolutionary concept of Conversational BI, using a WhatsApp bot to interact with BI Systems, focusing particularly on integrating with Google Analytics for E-Commerce Use Cases.

Also have a look at my GitHub Repository where I have build up a working showcase for Google Analytics.

Conversational BI - Talk to your data

Traditionally, BI relied heavily on dashboards and reports for data analysis and insights. However, thanks to Chat GPT, there’s a paradigm shift towards a conversational approach to reporting. Rather than shifting through graphs and charts, users can now simply converse with a chatbot to access the information they need.

With traditional BI, users rely on dashboards and reports for data analysis and insights.

With Conversational BI, users can chat with a bot to access the information they need. By incorporating the benefits of chatbots, such as chat history and full-text search in messages, users can seamlessly interact with their analytics data, receiving insights and reports directly within their messaging app.

Of course, it is also possbile to combine both approaches in one BI landscape.

Prompt Engineering

One fascinating aspect of this approach is that you don’t need to pass the data directly to the chatbot like ChatGPT. Instead, you describe the data model and query structure in the GPT Prompt, this enabling the bot to fetch the required information from the database.

Let’s take the example of engineering a Google Analytics API call.

I have build up a showcase in this GitHub Repository .

For generating the API Call, the ChatGPT prompt looks like this:

based on this query dimensions and metrics

Dimensions:
city
country
product
campaignName

Metrics
totalRevenue
totalUsers
bounceRate
sessions

create a JSON object like this:




based on this question:
{question}? 

if no time frame is given, the default time frame is the last 30 days

As a result, the Open AI Chat GPT Api will generate a query like this:

{
  "Dimension": "city",
  "Metric": "totalRevenue",
  "startDate": "2023-01-01",
  "endDate": "2023-12-31"
}

that can send be passed to the Google Analytics API to fetch the required data.

Benefits of Conversational BI

Imagine the convenience of receiving real-time analytics updates or generating customized reports—all through a simple chat interface. This conversational BI approach not only streamlines the reporting process but also enhances user engagement and accessibility to data.

This showcase combines the ChatGPT API, Google Analytics API, and WhatsApp API to build a communication channel that transforms the way businesses interact with their data. By integrating these technologies, users can effortlessly query their analytics data, share insights with team members, and stay informed on the go.

Obviously, this approach is not limited to Google Analytics. It can be applied to any data source not matter if it has an API or is a SQL Database.

Are you intrigued by the idea of implementing a reporting bot for your business? Consider the possibilities that Conversational BI offers in terms of data accessibility and user experience. Whether you’re a small startup or a large enterprise, embracing this innovative approach can revolutionize how you analyze and leverage your data.

Furthermore, there’s potential to enhance performance by fine-tuning the conversational model, ensuring more accurate insights and faster response times.

In conclusion, Conversational BI represents a significant leap forward in data analytics, making insights more accessible and actionable than ever before. Are you ready to chat with your data?

---