March 2026

March 18, 2026

Custom Time-to-Live (TTL) settings for trace retention

You can now configure custom TTL settings for trace spans in the AI Trust Layer. This update gives you full control over how long trace data remains available for review, allowing you to tailor retention to your privacy, compliance, and operational needs.

Through the AI Trust Layer policy in Automation Ops, you can:

• Enable or disable trace retention enforcement at the tenant level.
• Select how long traces are stored.
• Optionally restrict metadata sent to Insights to limit long-term data exposure.

Expired trace data is automatically removed based on your selected TTL, helping you manage storage, minimize sensitive data retention, and align with governance requirements.

This enhancement improves transparency and control over trace visibility across your automations.

For details, refer to Agent traces.

March 13, 2026

Unified runtime for agents

Unified runtime is now the default execution engine for agents, providing a single, Python-based runtime for both low-code agents (built with Agent Builder) and coded agents across design-time and production runs.

This includes:

• Improved consistency between debug and production runs
• A single tracing service captures runs across all agent types, while trace content and structure vary by agent type and framework

Existing agent workflows, build and debug experiences, and run results are unaffected.

For more information, see Running agents.

Note – March 16, 2026: Enablement is rolling out progressively across regions between March 13 and March 23. Not all regions are enabled simultaneously. If you notice any unexpected runtime behavior during this period, it may be related to the rollout reaching your region.

Guardrails observability in Agent Instance Management

The Guardrails tab is now available in Agent Instance Management, giving operators a centralized view of guardrail activity across deployed agents.

The tab includes:

• An Agents Guardrails Usage pie chart showing the proportion of agents with and without guardrails configured.
• A Top Invoked Actions stacked bar chart displaying the breakdown of Block, Escalate, and Filter actions per agent, filterable by top 5 or top 10.
• A Guardrails details table listing all deployed agents with their guardrail configuration and most recent enforcement data, including execution stage, action type, and action result.

For details, refer to Managing UiPath agents.

Note: The Guardrails tab in Agent Instance Management was released earlier but was not included in the release notes at the time.

March 9, 2026

Introducing data residency in the United Arab Emirates for Agents

As part of our ongoing cloud expansion, we are bringing complete data sovereignty to the United Arab Emirates region for Agents. This ensures your data is securely stored and replicated across multiple availability zones within the UAE, offering full control and compliance confidence.

For details, refer to Global cloud regions in the Automation Cloud Admin guide.