Process Routing

At enterprise volume, the bottleneck is rarely reading a document, it is getting each document to the right person and the right system without a human sorting the pile. Process Routing is how Staple directs documents automatically once they have been captured, so exceptions reach a reviewer, clean documents flow straight through, and every hop is recorded. A process in Staple is not a fixed pipeline you are locked into; it is a configurable graph you assemble for the document types you actually handle.

Built On A Graph Of Nodes

A Staple process is composed of nodes, each with a distinct job, connected into the path a document should follow. Trigger nodes bring documents in, for example a manual upload or an email drop. Execution nodes perform a discrete transformation, such as reshaping a JSON payload or evaluating an expression. Config nodes hold a setting once, for example a system prompt, so it can be reused across the process instead of redefined at every step. Service nodes do the heavy lifting, the extractor that reads a document, the data mapper that resolves it against your master data, and the reconciliation service that matches related documents. Webhook nodes fire an event when something happens, and HTTP nodes push data out to a destination. Because the process is a graph rather than a hard-coded sequence, adding a document type or a new branch is a configuration change, not an engineering project.

Ports Direct Each Document Type

Nodes connect through ports, and a port is typed to a specific document. A reconciliation node, for instance, accepts exactly the inputs its rule defines: if the rule names a purchase order as the anchor and an invoice as the target, the node exposes one port for the purchase order and one for the invoice, and each document type flows into the port built for it. Routing is therefore explicit and inspectable. You can see, for any document type, exactly which node consumes it and where it goes next.

Queue Rules Applied After Scanning

Staple applies queue rules to a document after it is scanned and classified. A rule can validate a field, set a value onto a document's sidebar field, or route the document onward, and rules are assigned per queue, so each document type follows its own path. A set-value rule can carry an optional set of conditions, so the value is written only when the document matches, for example stamping an entity code or cost center when a particular vendor is recognized. This builds on classification and splitting, which separates mixed document packs and preserves the relationship between related documents before routing begins.

Route By Content, Not By Filename

Documents are classified by what they contain, not by their filename, so an invoice, a delivery order, and a remittance in the same inbox are recognized and routed differently. Routing decisions can draw on extracted and enriched data, for example sending high-value invoices or a specific entity to a dedicated review queue, so the path a document takes reflects what is actually on it.

Exceptions To People, Clean Documents Straight Through

A document that fails a business rule or anomaly check, or whose fields fall below the confidence threshold, is routed to a reviewer with the reason attached, not passed downstream as if it were certain. A reconciliation that cannot be matched is held in the service until a missing document, such as a late credit note, arrives and the comparison can be rerun, rather than being force-fed onward. Documents that pass are routed to delivery without a manual touch, and role-based access controls who can view and act on each queue.

Every Move Is Recorded

Routing is auditable, not only its outcome. When a document moves from one queue to another, Staple records the source queue, the destination queue, and the user or rule that moved it, and it can fire a webhook carrying that event. A downstream system of record therefore sees the document's full path, not just where it finally landed. See how delivery events are consumed in Structured Data Delivery.

Configured Without Code

Queue rules and routing are configured in no-code configuration, so an operations lead can change how a document type is handled without waiting on a developer. Routing can also be driven programmatically through the Connect and Operate API, documented in the API documentation.

What Process Routing Does Not Do

Staple routes documents and their data to the right destination. It does not manage your organization's approval hierarchy or act as a case-management system, and it does not make the approval decision. It puts the right document in front of the right approver with the evidence to act.

Proven In Production

A Big 4 professional services firm processing around 10,000 documents per person per day for a global telecom client needed each document sorted, matched, and exceptions surfaced without a manual triage step. Staple classified and routed documents automatically and flagged discrepancies across documents for review, so people spent their time on the exceptions rather than on sorting the pile.

Related Capabilities

• Classification and Splitting

• Connect and Deliver

• No-Code Configuration

See Routing On Your Document Mix

Send us a representative inbox and we will show you how Staple classifies, validates, and routes each document type. Book a demo.