Databricks intros new tools to scale AI agents

Databricks launched new tools to help enterprises successfully deploy AI agents beyond the pilot phase into full-scale production.

The move addresses a significant challenge faced by many enterprises, as 85 percent of global enterprises now utilise generative AI (GenAI) yet struggle to achieve business-specific, accurate and well-governed outputs due to a lack of awareness of enterprise data.

The new suite of tools includes centralised governance for all AI models, allowing enterprises to integrate and manage both open-source and commercial models through the Mosaic AI Gateway. The gateway supports custom large language model (LLM) providers, providing unified governance, monitoring and integration across different AI models.

Also introduced is the AI/BI Genie Conversational API suite, which simplifies the integration of AI into existing workflows by enabling developers to embed natural language-based chatbots directly into custom-built apps or popular productivity tools such as Microsoft Teams, SharePoint and Slack. The Genie API is stateful, retaining context across multiple follow-up questions within a conversation thread.

To enhance human oversight, the Agent Evaluation Review App has been upgraded to streamline human-in-the-loop workflows. It is now easier for domain experts to provide targeted feedback and customise evaluation criteria without needing spreadsheets or custom-built applications, improving AI agent performance and accuracy.

Databricks has also unveiled Provision-Less Batch Inference that allows enterprises to run batch inference with Mosaic AI using a single SQL query. This eliminates the need to provision infrastructure, simplifying the integration of unstructured data and enabling seamless scalability.

“Many enterprises still struggle to deploy AI agents in high-value use cases due to concerns around accuracy, governance and security. The new tools address these challenges head-on, enabling businesses to move beyond pilots and into full-scale production with AI agents they can trust,” said Craig Wiley (top), Senior Director of Product for AI at Databricks.

Mosaic AI Gateway, Genie Conversation API suite, Agent Evaluation Review App, and Batch AI are now available in public preview.