July 2026 Summaries

In the Haystack Enterprise Platform, Traces provide comprehensive visibility into the execution of production agents by capturing the complete lifecycle of pipeline runs, from inputs to outputs, without requiring additional observability tools. This feature records every run's hierarchical component tree, token usage, logs, failures, and metadata, offering an interactive waterfall view for detailed analysis. Users can quickly identify issues, investigate latency, and understand component behavior, allowing for immediate reproduction and testing of fixes. Traces also facilitate comparison between pipeline versions, highlighting changes in latency, token consumption, and outputs, thus empowering users to assess improvements and costs accurately. Data security is maintained with role-based access controls and encrypted storage, ensuring sensitive information remains secure on deepset-owned infrastructure, with the option to export traces as needed. Available now in the platform, Traces automatically generate upon deploying or redeploying pipelines, and users can explore its capabilities through documentation or demos tailored to their use cases.

AI Sovereignty is increasingly discussed in the tech industry, with deepset emphasizing the importance of organizations building and managing AI systems independently to avoid vendor lock-in. This independence extends beyond deployment to the entire architecture, focusing on the accumulated knowledge and processes, known as institutional intelligence, surrounding AI models. Haystack 3.0 addresses these concerns by providing a framework that maintains control over the knowledge, logic, and environment of AI agents, supporting model independence and open-source operability. It introduces modular skills, agent hooks, and introspection features, allowing organizations to customize and control the AI's operational context, thereby ensuring that critical assets remain portable and under user control. This framework is especially trusted by regulated institutions due to its transparency and adaptability, which further enhances its suitability for enterprise deployment by allowing for easier testing, auditing, and integration management.