UiPath introduced more than a dozen updates to its automation platform at the FUSION conference in Las Vegas, led by a process-discovery tool called Cartographer. Cartographer is available from today, while upgrades to Test Cloud for software testing are scheduled for next month. The platform, used by more than 3 million developers, combines AI agents, scripts and third-party applications to handle repetitive business work. For buyers, the central promise is less manual preparation before automation reaches production.

UiPath adds Cartographer, testing and integration tools to platform

Cartographer maps work before automation starts

Cartographer addresses the first stage of automation projects: establishing how employees actually perform the manual task selected for acceleration. UiPath said this discovery phase can take weeks because teams must document both the standard procedure and responses to errors and unusual situations. Cartographer collects that operational data automatically and generates a ready-to-use automation workflow with built-in effectiveness measurement. Founder and Chief Executive Officer Daniel Dines said a conventional process map shows how work is designed to run, while the new tool captures how it runs in practice, including exceptions and judgment calls.

Once a generated workflow is deployed, a companion capability called Decision Ledger records the actions it performs in production. Enterprises can review that log to refine accuracy and keep the documented process synchronized with real execution. The approach replaces part of the interview-based mapping and manual workflow construction that typically precedes deployment. Measurement functions are included in the generated workflow itself, so teams do not assemble tracking separately. Together, discovery, generation and logging form a closed loop from observation to operation.

Test Cloud, the platform component for software testing, receives Autonomous Exploration next month. Engineers currently check applications with scripts that simulate user activity, and writing those scripts plus collecting results and extracting data points takes substantial effort. Autonomous Exploration lets engineers define high-level bug-detection goals instead of scripting each step. Test Cloud will also run manually created tests in automated mode, blending the two approaches, while a new AI agent called UiPath Delegate takes over administrative work such as preparing interface test summaries.

What this means for automation buyers

The third focus is Integration Service, which provides connectors between automation workflows and third-party applications. UiPath is adding connections to databases and to MCP-based AI agent tools, expanding the range of systems a workflow can read and update. A capability called Relay links workflows to on-premises workloads without moving those systems. Customers can keep workflow credentials for external applications in their existing secrets vaults rather than storing a separate copy inside UiPath, which simplifies administration and credential rotation.

Governance and developer tooling round out the release. New governance controls are intended to help enterprises keep the AI components inside workflows operating reliably in production. UiPath Coding Agents, first introduced in March, has moved to general availability and lets developers build automation workflows with AI coding tools such as Claude Code. For small firms, the practical effect is access to discovery, testing assistance and integrations without building separate scaffolding. For large organizations, the relevant change is centralized oversight combined with connections to distributed databases, on-premises systems and existing vaults.

The marker to watch is the Test Cloud rollout next month and how the first production logs from Cartographer and Decision Ledger perform. Adoption across the installed base of more than 3 million developers will show whether automatically generated workflows hold up against manually mapped processes. The decisive evidence will be documented accuracy improvements derived from ledger data and reduced rework in test-script maintenance. If those results appear, the discovery and testing bottlenecks described by UiPath will have narrowed in practice.