Databricks, a leading data infrastructure services provider, has announced a significant update to its Mosaic AI Agent Evaluation module with the introduction of a synthetic data generation API. This new API is designed to accelerate the development and testing of multi-agent systems, enabling enterprises to deploy these systems in production faster and more efficiently.

Multi-agent systems, also known as Agentic AI, have gained popularity among enterprises due to their ability to follow instructions, make decisions, and take actions independently, much like human workers. However, the development and testing of these systems can be a time-consuming and costly process. The new API from Databricks aims to address this challenge by providing a faster and more accurate way to evaluate agents.

The synthetic data generation API leverages an enterprise's proprietary data to generate evaluation datasets tailored to specific use cases. This approach eliminates the need for manual evaluation data-building, which can be time-consuming and prone to inaccuracies. By using synthetic data, developers can quickly generate evaluation data, reducing development costs and accelerating the testing process.

According to Databricks, its enterprise customers are already seeing the benefits of the new API. For instance, engineered components manufacturer Lippert has reported a 60% improvement in model response using synthetic data. This demonstrates the potential of the API to significantly enhance the development and deployment of multi-agent systems.

So, how does the API work? The process involves three steps: calling the API, specifying the number of questions, and setting natural language guidelines to assist synthetic generation. The API then generates a set of question-synthetic answer-source document groupings based on the enterprise data in the Agent Evaluation schema. Enterprises can then use the MLflow Evaluation UI to review the results of the quality analysis and make changes to agents to improve quality.

Databricks also offers enterprises the option to analyze the synthetic data generated by subject matter experts. This feature enables experts to quickly review the synthetically generated evaluation data for accuracy and add additional questions, ensuring that the data is accurate and relevant to the specific use case.

The introduction of this synthetic data generation API marks a significant milestone in the development of multi-agent systems. By providing a faster and more efficient way to evaluate agents, Databricks is enabling enterprises to deploy these systems in production faster, which can lead to improved productivity, reduced costs, and enhanced decision-making capabilities.

As the demand for multi-agent systems continues to grow, the new API from Databricks is poised to play a critical role in accelerating the development and deployment of these systems. With its ability to generate high-quality synthetic data, this API has the potential to transform the way enterprises approach agent development and testing, enabling them to unlock the full potential of Agentic AI.