In the fast-paced world of BI and Analytics, generative large language models (LLM) are revolutionizing the landscape. While these AI marvels have captured the public imagination with their ability to answer random queries and assist with homework, their true potential lies in transforming how businesses operate. As an analyst, you might be wondering: How can you harness the power of generative AI to redefine business analytics by streamlining your workflows, delivering unprecedented value to your organization?

Imagine a world where your analytics tool understands your intentions as easily as a seasoned colleague. Oracle previewed the AI Assistant for Oracle Analytics Cloud (OAC) at CloudWorld 2023. Demos showcasing a leap in analytics productivity left audiences in awe. This groundbreaking innovation centered on embedding AI and ML in OAC set the stage for Oracle’s recognition as a Leader in the 2024 Gartner® Magic Quadrant™ for Analytics and Business Intelligence Platforms, primarily for its pioneering embedding of AI and ML throughout the platform. The AI Assistant for OAC is now generally available as part of the September 2024 update.

Watch the Oracle Analytics bake off demo from the 2024 Gartner Data and Analytics Summit (12 min)

In addition to the AI assistant for analysts, new capabilities are coming soon for consumer users, allowing them to construct new dashboards using predefined templates, such as a sales dashboard. Users can easily drag and drop curated sales-oriented visualizations from the unified analytics catalog and add AI/ML-powered contextual recommendations without needing any prior understanding of which data sets or data connections to use.

The Oracle Analytics AI Assistant

Choice of built-in or bring-you-own language model

The embedded LLM is a specialized model designed to recognize the Oracle Analytics current workbook and datasets and understand the context of the user’s question within the current data focus. It is more accurate and less prone to hallucinations because, unlike external LLMs, it is optimized for analytics conversations and tasks.

An example command: “What are the top 5 countries by revenue for our gold level loyalty customers?“

While the embedded model excels at high accuracy for specific data questions, you may need more versatility for complex questions that require reasoning with both corporate and public data. This is where the bring-your-own-LLM option comes in.

The bring-your-own-LLM (BYO-LLM) capability in OAC allows you to integrate your own LLM subscription from services such as OpenAI. Language models from the outside world often have additional knowledge thanks to being trained on data about the outside world beyond the data sets in the analytics project, providing greater flexibility when analyzing complex scenarios. Using public data can potentially increase the risk of hallucinations by the LLM when using this option.

An example command: “What was the sales revenue on U.S. federal public holidays in 2023 for the top 10 most populous U.S. cities when the average temperature was 22°C or higher?”

In this example, the analytics project’s dataset comprises a revenue table, a date table, and a regional table. The additional information about U.S. public holidays, city populations, and historical weather data come from the LLM’s public training data. The LLM fills in these additional details, providing the correct information to accurately complete the query.

Availability

The AI Assistant is now available in all regions to customers using large OAC shapes. Over the next several months the AI Assistant will be steadily rolled out to all OAC customers. The AI Assistant will eventually be available to all OAC customers at no additional cost.

Key takeaways

The AI Assistant ushers in a new era of intuitive analytics by serving as your trusted partner. It reduces the need for intricate product knowledge and can enhance productivity by minimizing the amount of mouse operations needed to achieve the same outcomes. With a simple conversational command, an analyst can request “a six-period revenue forecast using Seasonal ARIMA on the line chart.” The AI Assistant translates natural language into actions, bridging the gap between an analyst’s vision and its realization on the canvas. It not only saves time but also empowers analysts to focus on insights rather than tool mastery. Moreover, it enables analysts experienced with other tools, such as Power BI or Tableau, to seamlessly transition to OAC without the steep learning curve typically associated with adopting new tools.

The top image illustrates the standard approach to building, which requires both knowledge of the product and data sets. In contrast, the bottom image shows the AI-assisted method for modifying the dashboard.
The top image illustrates the standard approach to building, which requires both knowledge of the product and data sets. In contrast, the bottom image shows the AI-assisted method for modifying the dashboard.

AI in analytics is not a distant promise—it’s here, and it’s transforming the way you work with data. The Oracle Analytics AI Assistant does more than help you keep up with technology—it fundamentally changes your approach and helps you gain a competitive edge. Organizations that embrace AI-driven analytics tools will be better positioned to make faster, more informed decisions in a data-driven world. Analytics powered by AI is conversational and intuitive—available to everyone now with the Oracle Analytics AI Assistant.

For more on Oracle Analytics: