UiPath has introduced a set of tools aimed at moving enterprise automation and AI agents from isolated deployments into governed business processes.
The announcements at FUSION 2026 include UiPath Cartographer, new agent-governance and connectivity controls, and planned Test Cloud capabilities for more autonomous software testing.
The common theme is that enterprise agents need context, permissions and testing around the work they perform rather than operating as standalone AI features.
Cartographer builds a map of enterprise work
UiPath Cartographer is designed to create and maintain a living map of how a business process actually runs, including exceptions and error-handling steps that are often missing from documentation.
The tool can generate process-design and solution-design documents and feed those process maps into UiPath Maestro for orchestration.
A Decision Ledger records decisions made during execution and can feed them back into the process map. UiPath’s aim is to make the operating context available to both people and agents rather than leaving it scattered across interviews, slide decks and outdated process documents.
More controls around agents and models
UiPath also announced governance features intended to keep agents and automations within enterprise-defined boundaries.
These include Model Hub, Runtime Checker, an LLM-as-Judge guardrail, compliance packs, and identity and access policies. Model Hub gives administrators a central view of which large language models are being used across UiPath products, including provider, model, routing region and lifecycle status.
That visibility matters as enterprises use multiple models and need to manage data residency, model changes and access policies across different automated workflows.
UiPath Coding Agents are now generally available, while UiPath Delegate is intended for administrative tasks. The company also expanded Integration Service and Data Fabric connectivity.
Testing moves toward autonomous operation
UiPath’s Test Cloud roadmap includes autonomous test execution and an Autonomous Exploration capability in which an AI agent investigates an application against a defined goal.
The company is also adding semantic and AI-based verification for applications that themselves contain AI features or agents, along with Playwright integration for teams already using that testing framework.
UiPath describes the longer-term direction as a ‘dark testing factory’, in which much of the testing system can operate autonomously rather than requiring people to write, run and maintain every individual test.
JFrog launches Traffic Controller to govern software package downloads by developers and AI agents

