Data fuels innovation, and innovation drives business success. Yet, the very business users who rely on data the most often face barriers—lacking the skills to query data, create visualizations, or uncover game-changing insights. This disconnect slows decision-making and impacts efficiency.
Imagine a world where business users no longer rely on traditional query models but instead engage in direct conversations with their data. A platform that eliminates complexity and latency in reporting, allowing users to simply ask for the insights they need—instantly.
Autonomous Database (ADB) Select AI bridges this gap by enabling natural language interactions with critical business data. The Select AI Sidecar uses an Oracle Autonomous Database instance (either a separate one or your existing ADB) that works alongside your other databases to offload SQL translation and federated queries. With a secure ADB AI sidecar repository, seamless integration with your preferred Large Language Model (LLM), and federated query capabilities across multiple clouds and data sources revolutionizes data accessibility.
Enhanced with Materialized Views and cost-effective storage, it ensures rapid, efficient data refresh. Now, business users can access insights without bottlenecks or delays, empowering them to make smarter, faster decisions with ease.
How does Select AI work?
Select AI enables users to query their data using natural language as an initial LLM prompt. Select AI enhances these prompts by incorporating schema metadata from your database, ensuring that the SQL generated by the LLM is accurate and reliable.
ADB Select AI supports showsql and runsql commands to display and execute the generated SQL, along with narrate and chat commands for a more conversational, interactive experience. By simplifying natural language queries, Select AI empowers non-technical business users to easily access and analyze data without requiring advanced SQL knowledge.
Figure 1. The original natural language query is augmented with metadata and context from database schemas to generate accurate SQL from the LLM
To learn more about the capabilities of Select AI please see Introducing Natural Language to SQL generation on Autonomous Database.
Data security with Select AI
Select AI can be deployed as a secure, fully managed AI sidecar database, efficiently translating natural language into SQL within a secure ADB instance. It enforces all database security features for data governance, including privileges, row-level security policies (VPD), and auditing, ensuring that Select AI-generated queries comply with established security standards. When integrated with Real Application Security (RAS), Select AI enables secure data access while maintaining robust database access controls and auditing.
Federated Queries with Select AI
The Select AI sidecar supports federated queries across diverse datasets, whether on-premises or in multi-cloud environments, leveraging the Select AI pipeline. ADB Select AI enables non-technical users to generate SQL for complex federated queries, streamlining execution, eliminating data duplication and ETL processes, and ensuring more accurate results.
For example, a natural language query like “Show pending orders for Acme Corp” can seamlessly retrieve customer data from a BigQuery table on Google Cloud and order details from an Amazon Redshift database on AWS. Select AI handles the complexity of joins, data locations, and query optimization—no need for users to write SQL, know data locations, or move data.
With a fully managed database and Real Application Security (RAS) for end-to-end security, Sidecar ADB Select AI provides the essential foundation for successful AI-driven data access and analysis.
Figure 2. ADB Select AI Sidecar Architecture and Data Flow leveraging federated queries.
Sidecar Select AI with Retrieval Augmented Generation (RAG)
Select AI already simplifies and automates using generative AI in combination, whether generating, running, and explaining SQL from a natural language prompt or chatting with the LLM. Now with RAG, you can take your use of generative AI to the next level, conveniently from SQL:
- Create and populate vector stores using your private data via the built-in automated AI pipeline
- Automatically augment your prompt or query with vector store content retrieved through semantic similarity search
- Send this augmented prompt to your specified AI provider and LLM and returning generated results
Select AI supports RAG using Oracle AI Vector Search, which is a new feature of Oracle Database 23ai.
Seamless Integration with Oracle Tools
Select AI seamlessly integrates with a variety of tools, including SQLcl, SQL Developer Web, Oracle Application Express (APEX), and Oracle Machine Learning Notebooks. This broad compatibility allows users to incorporate Select AI into their existing workflows, enabling effortless data querying and analysis.
For instance, an APEX application can connect with a Select AI sidecar and OCI Generative AI to facilitate natural language interactions with an Oracle HCM system. The following demo showcases the Ask Oracle E-Business Suite (EBS) APEX UI, which interfaces with an ADB sidecar instance to generate augmented SQL prompts. These prompts are then processed by OCI Generative AI LLM. In this demo, a conversation unfolds with three sequential questions related to employee qualifications, each drilling down for more detailed insights—all expressed through natural language queries.
Figure 3. Ask Oracle E-Business Suite APEX app with ADB Select AI Sidecar and OCI Generative AI LLM .
Summary
ADB with SelectAI transforms data accessibility, empowering business users to query complex, distributed data effortlessly through natural language conversations. Select AI bridges the gap between users and their data, accelerating innovation and simplifying data interactions.
As a seamless AI sidecar, Select AI offers an easy entry point into your generative AI journey. Ready to dive in? Start by deploying an ADB Free Tier instance, explore Select AI’s capabilities, and experiment with its powerful federated query features today!
Check out the references below to get started!
References
Conversations are the next generation in natural language queries
How to help AI models generate better natural language queries
Autonomous Database Select AI: Accelerate innovation with enterprise data, OCI Generative AI and enhanced security
Announcing Select AI with Retrieval Augmented Generation (RAG) on Autonomous Database
Enabling natural language query of EBS 122 using Oracle generative AI
Introducing natural language to SQL generation on autonomous database
Select AI Livelab workshop
Oracle Database Real Application Security
DBMS_CLOUD_AI
Authors: Kumar Rajamani, John McHugh, and special thanks to Nadia Benjedou for insights on the EBS ADB Select AI deployment.