Elastic has announced the release of Agent Builder, a suite designed to assist developers in creating secure, context-driven AI agents. This tool, built on Elasticsearch, supports context engineering, providing a platform that scales, searches, and analyses enterprise data.
The introduction of Agent Builder aims to address the need for context in AI tasks, simplifying agent workflows through native data preparation, retrieval, ranking, and observability. Developers can interact with their data or create a custom context-driven agent in a short amount of time.
One intended advantage of Agent Builder is its support for MCP and A2A protocols, as highlighted by Amanda Silver, CVP, Microsoft CoreAI. This integration allows deployment with Microsoft’s frameworks, using Elasticsearch as a knowledge source.
Elastic's new tool also reduces complexities faced in the current landscape of AI development. Sam Partee, Arcade.dev’s co-founder, emphasised that the partnership with Elastic aims to offer developers a secure platform to advance agents from demonstration to production-ready stages for structured context retrieval and reasoning.
Jerry Liu, CEO of LlamaIndex, highlighted the significance of unlocking enterprise context from diverse data sources. The collaboration with LlamaIndex seeks to enhance document processing capabilities, strengthening the agent's ability to process and reason with accuracy.
In conjunction with Agent Builder, Elastic has introduced the Elastic Workflows in tech preview. This feature aims to extend the agent building's capabilities by allowing agents to take autonomous actions across systems, filling the gap left by conventional AI planning. Workflows orchestrates internal and external systems, enabling precise data management and action execution.
Ken Exner, Elastic's chief product officer, remarks on how Agent Builder and Workflows collectively seek to provide a secure, reliable platform, important for enterprises to transition successfully from pilot phases to substantial real-world impacts.
The architecture of Agent Builder is designed to be versatile, being model-agnostic and compatible with model-as-a-service providers, including leading cloud hyperscalers. This compatibility supports broad usage, enabling its application in AI agent development for enterprises.
