Blue Prism Technical Architect key terms

The top-level Blue Prism artifact that models an end-to-end business workflow, built from stages in Process Studio and responsible for calling business objects to interact with applications.

A reusable Blue Prism component that encapsulates all interactions with one application or system, exposing Actions that processes call instead of automating the application directly.

A published, callable entry point on a business object that performs one discrete task, such as logging in or reading a record, and can be invoked from a process or another object.

A tabbed subdivision of a process or object that groups related stages, supports reuse through page references, and keeps logic readable and testable.

The Blue Prism component used to spy and interrogate application user interfaces, capturing elements and attributes that become the Application Model an object automates against.

The stored hierarchy of windows and elements that represents an application's interface inside a business object, enabling stable identification of controls at runtime.

A capture technique in Application Modeler (such as accessibility, HTML, or region modes) that determines which attributes Blue Prism records to identify and manipulate interface elements.

The named properties captured for an element, such as path, name, or class, which Blue Prism matches at runtime to locate the correct control on screen.

A single identified user interface control within an application model, with its own attributes, that actions read from, write to, or click on during automation.

The technique of replacing volatile parts of an element attribute with wildcard characters so identification survives values that change between sessions, such as dynamic IDs or timestamps.

A typed variable declared in a process or object that holds a value during execution, with a defined data type, initial value, and scope of visibility.

A tabular data structure in Blue Prism holding rows and named fields, commonly used to pass queues of work items or multi-record results between pages and objects.

The declared category of a data item, such as text, number, date, flag, or binary, which constrains the values it can hold and the operations applied to it.

A named configuration value stored outside process logic, letting the same release run against different environments by changing settings rather than editing and republishing processes.

A Blue Prism store of work items with statuses such as pending, locked, and completed, enabling multiple robots to share workload, retry failures, and report throughput.

A single record in a work queue carrying the data and status needed to process one unit of work, including attempts, priority, tags, and exception details.

A queue item deliberately postponed until a specified future time, used when a dependent system or condition is not yet ready and immediate retry would fail.

The structured use of recover, resume, and block stages to detect errors, decide whether to retry, skip, or fail, and record clear diagnostic information.

A stage that marks the start of a block whose errors are caught rather than propagated, giving the developer a defined point to handle failure and choose the next step.

A stage paired with a recover block that defines where execution continues after an error is handled, typically retrying the attempt or moving to cleanup logic.

An error raised when automation cannot continue, such as an application crash or missing element, which typically marks the queue item as failed and triggers retry or escalation.

A rule-based outcome where the data or situation is invalid rather than the automation broken, usually flagged so the item is not retried but routed for human review.

A secret such as a username, password, or token stored in the Blue Prism credential store and retrieved by name at runtime, so secrets never appear in process logic or logs.

The secured repository, often backed by a database or vault, that holds encrypted credentials and answers retrieval requests only for authorized users and robots.

The assignment of permissions to roles rather than individuals, controlling which users can view, edit, run, or release processes, objects, queues, and credentials.

The recorded history of changes and actions inside Blue Prism, supporting traceability of who changed what and when for governance and investigation.

A machine registered with Blue Prism that executes processes under a robot account, characterized by capacity, resource pool membership, and availability for scheduling.

A grouping of runtime resources that share work, allowing the scheduler to allocate queue items to whichever robot in the pool is free.

A single execution instance of a process on a runtime resource, with its own logs and status, which administrators monitor and can terminate when stuck.

The detailed record of stages executed during a session, including data values and errors, used for debugging and evidencing what the robot did.

The Blue Prism facility that starts tasks at defined times or intervals on selected resources or pools, forming the backbone of unattended overnight operations.

A packaged, versioned snapshot of processes, objects, and related components that is imported into another environment as the unit of promotion and rollback.

The Blue Prism component that builds release packages, tracks their contents and versions, and imports them into target environments.

Blue Prism's delivery and governance framework covering roles, processes, and controls needed to run an automation programme reliably from idea through live operation.

The central team that owns standards, tooling, pipeline governance, and support for automation, enabling consistent delivery and reuse across business units.

The agreed specification of a process to automate, capturing steps, systems, data, exceptions, and volumes, and serving as the baseline for build and testing.

The document translating requirements into an automation blueprint, defining process decomposition, objects, queues, error strategy, and infrastructure before build begins.

The practice of splitting a large business process into smaller processes, objects, and pages so work can be parallelized, reused, and maintained independently.

The configured behavior of re-attempting a failed queue item, bounded by a maximum attempt count, to absorb transient errors without manual intervention.

The rate at which completed work items are produced, used with item handling time and resource count to size robot capacity against business demand.

The ability to add runtime resources, split queues, or shard workloads so automation volume grows without redesigning processes or breaching service levels.