The Oracle Analytics AI Assistant

Oracle Analytics have just announced the general availability of the AI Assistant. It is a new feature that allows users to interact with their data using natural language. This means you can ask questions in plain English and the assistant will generate visualisations and insights based on your dataset.

Democratisation of Data Analytics: The AI assistant makes data analysis accessible to a wider range of users, not just those with advanced technical skills or even those with a formidable understanding of the data visualisation tool. By allowing users to interact with data using natural language, it eliminates the need for complex SQL queries or programming knowledge. This democratisation empowers more people within an organisation to explore data and derive valuable insights.

This feature is part of Oracle’s broader AI-driven efforts within Oracle Analytics Cloud, which also includes integration with machine learning models and advanced natural language generation for smart data narratives

Below, is a summary of the steps to follow to get the AI Assistant up and running:
​

• Enable your dataset: Ensure your dataset is ready for use by checking the Insights panel and enabling it if necessary.
• Train the assistant: Index your dataset to help the assistant understand your data and answer questions accurately.
• Ask questions: Start asking questions using natural language. The assistant will generate visualizations and provide insights based on your data.

This Oracle Analytics blog provides detailed information of these simple steps needed to set this up. It requires some configuration to enable the feature at a dataset level. To improve the user experience, setting up synonyms for the dataset attributes is useful as this gives you the option to provide alternative names for each attribute which will improve the results generated by the AI Assistant. 

Remember to choose the option to index the dataset for the Assistant. There is also an option to index the dataset for the Assistant and the Homepage Ask feature if required. 

Upon initial review, the Oracle Analytics AI Assistant shows significant potential for enhancing productivity and easing the workflow, particularly for those less familiar with advanced analytics tools. While the setup, particularly dataset configuration and indexing, may require some upfront effort, once configured, the tool truly shines. The integration of natural language querying makes it much easier to interact with the data, reducing the need to master the complexities of Oracle Analytics Cloud (OAC).

The feature’s ability to generate instant visual insights from natural language inputs is particularly valuable. This means that once your dataset is indexed and ready, you can quickly derive meaningful insights, spot trends, or explore relationships in the data without having to dive into manual visualisation processes. Auto Insights can also uncover patterns you may not have initially considered, adding further value by highlighting hidden connections.

Subject Areas
This version of the AI Assistant does not currently have the capability to support subject areas. However, this feature is part of the planned future enhancements and is expected to be integrated into the Assistant in due course. Once implemented, it will allow users to engage with the Assistant more effectively, offering deeper insights and tailored assistance based on specific subject areas. 

Oracle Analytics AI Assistant is a noteworthy addition to the Oracle Analytics platform. It offers a more intuitive approach to data analysis by allowing users to interact with data using natural language. While this feature can simplify certain tasks and provide valuable insights, it's important to note that it may not replace the need for human expertise in all cases. As with any AI-powered tool, it's crucial to understand its limitations and use it in conjunction with traditional data analysis methods.

Update - 19th September 2024

After some conversations with the Product Management team , I have gathered some more information about the AI Assistant pertaining to the rollout process and some details about security and data privacy which you can find below:.

Oracle’s objective is to make the AI Assistant available at no additional cost across all OCPU shapes. However, its release will be managed through a phased, controlled rollout that will span several Oracle Analytics Cloud updates. This approach ensures optimal performance and stability while addressing key concerns around privacy and security.

The Large Language Model (LLM) powering the Assistant is highly resource-intensive and Oracle-managed. Due to its proprietary nature and status as protected intellectual property, specific details regarding the LLM are limited. However, it is important to note that each Oracle Analytics Cloud instance is provisioned with its own private, Oracle-managed LLM.

To further safeguard data privacy, all interactions between the LLM and each individual Oracle Analytics Cloud instance remain entirely within the customer’s OCI environment. At no point do these interactions leave the customer’s infrastructure, ensuring complete data security and confidentiality.