Conversing with Data: The Impact of Generative AI on Business Intelligence
Conversational business intelligence (BI) marks the latest effort to break the analytics bottleneck and help companies make more data-driven decisions. This emerging technology uses generative artificial intelligence (GenAI) to guide and automate the various tasks of business intelligence through a chatbot. Most major BI tools now include conversational BI chatbots; an Eckerson Group survey shows a third of data practitioners already use them. When implemented well, conversational BI can make power users (i.e., data analysts and scientists) more productive and help casual users (i.e., data consumers and explorers) serve themselves.
GenAI does introduce risks—including accuracy, topic drift, and compliance—that can undermine the value of conversational BI. To overcome these risks and deliver the benefits of productivity and self-service, a conversational BI tool must be intuitive, accurate, consistent, secure, and open. Tools that meet these requirements support the following use cases:
> For power users, they guide code development, generate and debug code, explain and document code, generate schema descriptions, evaluate GenAI outputs, and create predictive features.
> For casual users, they recommend reports, generate insights, explain insights, explain predictive models, and activate functions.
Eckerson Group recommends that companies take these steps to address the opportunity of conversational BI:
> Start with power users. Data analysts and data scientists should be your primary early adopters because they have the motivation and knowledge needed to break the analytics bottleneck.
> Move on to casual users. Once your power users have experience and confidence with conversational BI, they can identify tasks to push back to casual users.
> Govern. Govern. Govern. Train power users and casual users alike to carefully inspect the outputs of these chatbots and help enforce data governance policies.