Cartographer is aimed at a problem that becomes more important as companies give AI systems a bigger role in day-to-day operations: much of the knowledge behind a business process still lives with the people doing the work. 

[For more news, click here]

A company can have detailed procedures for handling a claim, approving an invoice, or processing a loan and still find that the documents do not explain how the work actually happens. Employees make decisions based on experience, deal with exceptions that were never included in the original process and know when a case needs to be treated differently. Those details may be familiar to the people involved but remain scattered across training material, spreadsheets, conversations, and individual knowledge. The problem becomes more practical when the same work is handed to software that is expected to make decisions or take actions within the process. 

UiPath is addressing that issue with Cartographer, announced in September at its FUSION conference in Las Vegas. The product is designed to document a business process as it operates today, help teams design how it should operate in the future, and produce the documentation needed to build automation from it. UiPath calls the longer-term record a "Map of Work," describing it as a versioned and governed definition of how a business process runs. The company says Cartographer can take existing SOPs, diagrams, recordings, and other material as input, then work through a guided discovery process to establish the steps, business rules, decisions, and gaps that need to be resolved.  

The useful part of the approach is its attention to what happens outside the standard workflow. A process map can show the normal sequence of events, but difficult cases are often the ones that require someone to interpret a rule or decide whether an exception should be escalated. Cartographer's current workflow asks users to confirm an "AS-IS" process, including its business rules and gaps, before moving to a future-state design. Each step can then be assigned an approach such as human work, automation, or agent involvement, with the approved design used to generate a Process Design Document and, later, a Solution Design Document for development teams and coding agents.  

The product comes as AI systems are being given access to applications and business information where a mistake can have consequences beyond producing a poor answer. Companies need to define not only what an agent can access but also what it should do when a process reaches an unusual case. Process documentation has always mattered to automation projects, but the arrival of agents makes the quality of that documentation more consequential because the software may be acting on the information rather than simply helping an employee interpret it. 

"For years, businesses have digitized their data without ever capturing the enterprise context of how that data is actually put to work," said Daniel Dines, Founder and CEO of UiPath. "Map of Work is how we translate that gap into something concrete. A process map tells you how work is designed to run. UiPath Cartographer is the key to uncovering the operational knowledge of how that work actually runs, with every exception and every judgment call included. It’s the first product built to capture both and keep them in sync."  

Keeping the record current is part of the product's design. Cartographer is intended to maintain a versioned process definition rather than produce a document that becomes outdated once an automation project is complete. Its documentation also describes a workflow in which decisions made during the running of a process can be considered as proposed changes, with approval required before they become part of the governed process. The company's longer-term aim is to keep the process definition, sign-off, and resulting development documents connected rather than allowing each to become a separate version of how the business works.  

The challenge for customers will be maintaining that discipline once the initial mapping exercise is over. Policies change, applications are replaced, and employees find new ways to handle recurring situations, all of which can leave formal documentation behind. Cartographer does not remove the need for security controls or human oversight, but it addresses a basic requirement that can easily be missed in discussions about AI agents: before software can reliably perform a complicated business process, the organization needs a clear and current account of what that process involves and where human judgment still belongs. 

As enterprises move more work into software, the informal knowledge that once stayed with employees has to become easier to record, review, and maintain. Cartographer is UiPath's attempt to turn some of that knowledge into a working record that can be used by both people and the systems they are asking to carry out the job. 

Related Articles 

AI Agent Governance Is Falling Behind Enterprise Adoption, Optro Research Finds 

Red Hat Wants AI Agents to Prove Their Work Before They Get the Keys to the Business 

With AI Agents Up 460% in a Year, BeyondTrust Launches a Governance Tool to Match