Sphere builds its AI platform using Oracle Autonomous Database 23ai

Through its AI platform OrgBrain, Sphere Global Holdings provides AI-driven services for compliance, governance, and operations that support enterprise decision-making for complex organizations, including multinational banks, government agencies, and nongovernmental organizations. As a startup with a small development team, the company needed a cost-effective and high performance data platform that could handle massive data sets, integrate AI capabilities, and meet clients’ stringent data residency requirements. Using Oracle Database 23ai on Oracle Autonomous Database, Sphere created an AI platform using a single database that can draw on different data formats, including relational, vector, spatial, JSON, and graph, to dramatically simplify complex tasks, such as writing evidence-based grant proposals in a matter of minutes instead of weeks.

Sphere saw Oracle Autonomous Database 23ai on OCI as the right data platform upon which to build and run its sophisticated AI offerings. Autonomous Database 23ai offered Sphere unique advantages, including capabilities for storing and searching billions of vectors, natively integrating large language models, and handling a range of data formats. In the process of building the OrgBrain platform, Sphere evaluated several competing database approaches that would have required bringing in multiple specialty databases—one for vector, another for spatial, another for graph—and integrating them together. In contrast, Oracle Autonomous Database 23ai’s converged database approach eliminated the complexity of managing separate databases while meeting Sphere's scalability and performance goals.

Autonomous Database 23ai also stood out for its security and scalability, as well as for facilitating high performance, even with large data sets. Autonomous Database 23ai is designed to handle enterprise-scale data volumes and its performance benchmarks reflect the large size of real-world vector workloads, says Christian M. Grupp, Sphere Holdings CEO.

Using Oracle Autonomous Database 23ai helped Sphere quickly launch its AI-powered OrgBrain platform and pursue contracts with large, security-conscious organizations that require high-speed, scalable, and secure data services. After the company simplified its data infrastructure by consolidating multiple data types and storage needs into Autonomous Database 23ai, it no longer needed to integrate multiple databases. The company had just one security architecture for all its data. Because Autonomous Database 23ai supports all data types needed for OrgBrain in a single converged data platform rather than multiple solutions, Sphere was able to keep its startup development team small while creating a highly efficient and performant AI system for enterprise use.

Autonomous Database 23ai helped deliver the fast performance essential to Sphere’s platform. The database’s AI Vector Search capability processes thousands of substeps in queries, reducing processing time from minutes to seconds. This performance helped the company dramatically compress complex enterprise workflows for end users. “There’s no other way to say it—it's just fast,” says Grupp. “And that performance really is important when you start actually rolling this out and doing things.” For example, government agencies using Sphere could cut the time to create a first draft of a notice of funding opportunity from a week or two to about five minutes.

Autonomous Database 23ai also provides strong security and data residency features critical for Sphere’s clients, who work with healthcare, finance, and other regulated data. Oracle’s database lets the company manage access and permissions at the core data layer, helping to keep users’ visibility restricted to only the data they're authorized to access.

The company uses Autonomous Database 23ai as a central repository that connects and manages different enterprise knowledge sources and data and file types. This allowed Sphere to efficiently integrate, store, and retrieve diverse source content such as documents, audio, video, and more as part of its platform. Oracle Autonomous Database 23ai helped the company build a platform to handle multiple data types seamlessly and keep all information interconnected and accessible, which is key to OrgBrain’s ability to support decision-making.

Autonomous Database 23ai also gave Sphere’s developer team the built-in tools and capabilities they needed to be creative and move at a fast startup pace while knowing the resulting application will meet enterprise standards once in production. For example, JSON Relational Duality Views in Autonomous Database 23ai stores data in an efficient relational format but also makes it available in the JSON format, which is often favored by developers. Oracle REST Data Services saved the company months of development time by quickly mapping its data models to objects in Oracle Autonomous Database. Its integrated security model also eliminated the need for bolt-on authentication and data access controls for each new application, which reduced development complexity for Sphere’s lean team.