AA Master RPA cheat sheet

Control Room is the central web-based management console for users, roles, bots, devices, and credentials; it does not execute bot logic itself.

Bot Creator is the licensed component used to design and edit bots; Bot Runner executes bots and consumes runtime licenses.

A Control Room user can hold multiple roles; effective permissions are the union of privileges granted by all assigned roles.

The repository stores bot files with version history; check-in/check-out prevents concurrent edits by multiple users.

Devices (Bot Runners) register with Control Room via a unique device key or manual connection; a device can be online, offline, or disconnected.

Credential Vault stores encrypted credentials and lockers; bots retrieve credentials at runtime without exposing plaintext in the bot logic.

Automation Anywhere supports attended (triggered by user action) and unattended (scheduled or queued) bot execution modes.

The audit log records user and bot actions with timestamps; retention is configurable by the Control Room administrator.

Bot Insight provides operational and business analytics dashboards derived from bot execution data and custom variables.

Actions are the building blocks of a bot; each action has input parameters, output variables, and optional error-handling settings.

Variables must be declared with a type (string, number, boolean, list, dictionary, record) before use; type mismatches cause runtime errors.

The Recorder captures UI interactions into actions; recorded actions should be reviewed and parameterized for reuse.

Selectors identify UI elements; prefer stable attributes (name, class, control type) over screen coordinates for resilience.

The If/Else, Loop, and Error Handler actions control flow; nested loops increase complexity and should be limited.

String actions include split, replace, substring, trim, and regex matching; use them to parse text instead of hardcoding values.

The Log To File and Message Box actions aid debugging; production bots should log to a central file or Control Room.

The Delay/Wait action synchronizes with slow applications; prefer Wait For Condition over fixed delays where available.

Bot export/import packages (.atmx or .bot) include dependencies; verify version compatibility before importing to another Control Room.

Use a parent-child bot architecture: a parent bot orchestrates child bots via the Run Task action and passes variables as input/output parameters.

Error Handler actions (Try, Catch, Finally) isolate failure-prone steps; Catch blocks should log the error and decide whether to retry, continue, or terminate.

The On Error action can be set to resume next, go to a label, or fail the bot; choose based on whether the step is critical.

Implement retry logic with a counter and delay for transient failures (network timeouts, locked files); cap retries to avoid infinite loops.

Use a centralized logging convention (timestamp, bot name, step, severity, message) to simplify troubleshooting across child bots.

Store configuration values (URLs, paths, thresholds) in a config file or Control Room credential/variable store, not hardcoded in actions.

Idempotency: design bots so re-running a failed instance does not duplicate transactions; use unique keys and status checks.

Queue-based processing: use Work Items (queue) with status (new, in progress, completed, failed) to distribute work across Bot Runners.

Handle application pop-ups and session timeouts with dedicated error handlers and recovery subroutines that re-login and resume.

Deployment lifecycle: develop in Bot Creator, check in to repository, deploy to Bot Runner devices, then schedule or trigger execution.

Schedules define when bots run (once, daily, weekly, cron-like); they are assigned to devices or device pools.

Device pools group Bot Runners for load balancing; queued tasks are distributed to available devices in the pool.

Workload management (queues) lets you enqueue work items and assign them to bots; monitor queue depth and failure rates.

Role-based access control: assign least-privilege roles (e.g., Bot Developer, Bot Runner, Administrator) to users and groups.

Version control: use check-in with comments; roll back to a prior version if a deployment fails validation.

Deployment checklist: validate bot dependencies, credentials, device availability, schedule conflicts, and rollback plan before go-live.

Control Room dashboards show bot status (running, completed, failed), device health, and queue metrics for operational monitoring.

Audit and compliance: export audit logs and execution history for change management and regulatory reviews.

The Center of Excellence (CoE) defines standards, reusable components, review gates, and training for bot development.

Least privilege: grant users and bots only the permissions and credentials required for their tasks; review access quarterly.

Credential Vault must be used for secrets; never store passwords in bot variables, config files, or logs.

Change management: require peer review, testing evidence, and approval before promoting bots to production.

Segregation of duties: developers should not have production deployment rights without a separate approver.

Data protection: classify data handled by bots; mask or exclude sensitive data from logs and analytics.

Bot lifecycle governance: maintain an inventory with owner, purpose, risk rating, and review date; retire unused bots.

Incident response: define escalation paths for bot failures, including on-call ownership and rollback procedures.

CoE metrics: track bot utilization, success rate, hours saved, and exception rates to prioritize improvements.