In a workflow automation, what starts an event-driven flow?
A A chart legend
B A dashboard color
C A trigger event
D A completed report
Answer: C. A trigger event. A trigger defines the event or condition that starts an automated workflow.
What does an action represent in a workflow?
A A task the flow performs
B The event that starts it
C A user password
D A stored browser tab
Answer: A. A task the flow performs. An action is a step the workflow performs after its trigger.
How does a scheduled automation differ from an event-triggered one?
A It always needs approval
B It cannot use actions
C It changes source data
D It runs at set times
Answer: D. It runs at set times. A scheduled flow runs at configured times, while an event-driven flow starts when a specified event occurs.
Why add a trigger condition before launching a workflow?
A Hide the event history
B Avoid unnecessary flow runs
C Increase duplicate actions
D Remove all error handling
Answer: B. Avoid unnecessary flow runs. Filtering at the trigger can reduce runs for events that do not meet the workflow's conditions.
What does a condition step allow a workflow to do?
A Change its trigger type
B Guarantee input accuracy
C Disable permissions
D Choose a data-based path
Answer: D. Choose a data-based path. A condition evaluates data and directs the flow along different branches.
When should workflow steps run in parallel?
A When order is essential
B When one needs another's output
C When they are independent
D When the source is unknown
Answer: C. When they are independent. Independent actions can run in parallel; dependent actions must wait for needed results.
Why should a workflow define its step order?
A Every action is simultaneous
B Later steps need earlier outputs
C Triggers need no conditions
D Logs require a sequence
Answer: B. Later steps need earlier outputs. The order matters when a later action depends on data or status produced earlier.
Which workflow step adds human judgment to an approval process?
A A sign-off request
B A retry delay
C A file rename
D A date filter
Answer: A. A sign-off request. An approval step routes a decision to a person before later actions proceed.
What is the risk of retrying a non-idempotent action?
A It may repeat side effects
B It improves source accuracy
C It removes all duplicates
D It changes the trigger
Answer: A. It may repeat side effects. If an action is repeated after a timeout, it may create duplicate payments, messages, or records unless duplication is handled.
How can a workflow reduce duplicate processing of the same event?
A Retry without limits
B Remove execution logs
C Disable trigger filters
D Use a unique event key
Answer: D. Use a unique event key. A unique event identifier can help detect and ignore events already processed.
Which failure is most suitable for a limited retry?
A An invalid required field
B A temporary service outage
C A user authorization denial
D A permanently missing record
Answer: B. A temporary service outage. Retries can help with transient failures, while permanent validation or permission errors need correction.
Why use a backoff delay between retries?
A Increase invalid input
B Hide failures
C Reduce repeated pressure on services
D Change event identity
Answer: C. Reduce repeated pressure on services. Increasing delays can reduce pressure on a service that is temporarily unavailable.
What can catch events that fail after the retry limit?
A A chart label
B A new trigger
C A hidden worksheet
D A failure queue or alert
Answer: D. A failure queue or alert. A failure queue or alert provides a place to inspect events that could not be processed.
Which workflow information helps diagnose a failed run?
A Only the flow's color
B The user's wallpaper
C Step status and error details
D A random prompt
Answer: C. Step status and error details. Run history, step status, and error details help locate the point and cause of failure.
Why include an error-handling path in an automation?
A Handle step failures safely
B Guarantee no failures
C Remove required permissions
D Change business rules
Answer: A. Handle step failures safely. An error path can stop, notify, retry, or route work for review instead of continuing blindly.
What should happen after a critical step fails?
A Continue as if it succeeded
B Stop dependent actions
C Delete the failure log
D Repeat every earlier step
Answer: B. Stop dependent actions. Dependent steps should not proceed as though a failed prerequisite succeeded.
Why set a timeout on a long-running action?
A Guarantee task success
B Prevent indefinite waiting
C Hide service outages
D Change event order
Answer: B. Prevent indefinite waiting. Timeouts bound how long the flow waits and allow it to handle a stalled service.
Which signal can show that an automation is silently failing?
A A different theme
B A longer flow name
C Extra dashboard colors
D Repeated failed-run alerts
Answer: D. Repeated failed-run alerts. Monitoring failure rates and alerts can reveal problems that are otherwise not visible to users.
How should an automation communicate a completed action?
A Report its actual result
B Claim success after failure
C Hide skipped steps
D Invent missing data
Answer: A. Report its actual result. Status messages should reflect what the workflow actually completed, skipped, or failed.
Why monitor workflow volume as well as failures?
A Volume proves accuracy
B Low volume proves success
C Unexpected spikes can signal loops
D Spikes remove rate limits
Answer: C. Unexpected spikes can signal loops. A sudden increase in runs can indicate a repeated trigger or unintended feedback loop.
What can prevent an automation from repeatedly triggering itself?
A Remove all conditions
B Increase retry attempts
C Share broader access
D Exclude its own updates
Answer: D. Exclude its own updates. Trigger conditions can filter the workflow's own updates if those updates would otherwise cause another run.
Which control helps an automation respect a service's rate limit?
A Send all requests at once
B Throttle or pace requests
C Retry immediately forever
D Ignore limit responses
Answer: B. Throttle or pace requests. Throttling spreads requests and reduces the chance of exceeding service limits.
How should a workflow's credentials be scoped?
A Administrator access everywhere
B A shared public password
C Only permissions it needs
D No expiration or review
Answer: C. Only permissions it needs. Least-privilege access limits the data and actions available to the workflow.
What should a workflow owner do when a connector account changes?
A Review authentication and access
B Share credentials in logs
C Disable all audit records
D Ignore failed connections
Answer: A. Review authentication and access. Connection changes can affect access and ownership; review authentication and permissions after updates.
Why should an automation collect only data needed for its task?
A Increase storage use
B Make prompts longer
C Reduce unnecessary exposure
D Create more branches
Answer: C. Reduce unnecessary exposure. Data minimization reduces the amount of information flowing through connectors and logs.
What is useful before enabling a new automation for everyone?
A Run it on all records
B Test it on sample cases
C Disable monitoring
D Skip approval checks
Answer: B. Test it on sample cases. A controlled test helps reveal incorrect conditions or actions before the workflow affects live records.
Where is a safer place to test a workflow that changes important records?
A The production database
B A public folder
C A shared admin account
D A test environment
Answer: D. A test environment. A test environment helps verify behavior without changing live business records.
Why should an automation test include boundary cases?
A Conditions may fail near limits
B Boundaries change file names
C It removes error handling
D All inputs are identical
Answer: A. Conditions may fail near limits. Values near thresholds can expose mistakes in condition logic or range handling.
When should a workflow use a compensating action?
A To address a partial completion
B Before its trigger fires
C To hide every error
D When nothing changed
Answer: A. To address a partial completion. A compensating action can help restore a consistent state when some steps succeed and a later step fails.
What should be checked before an automation sends an external message?
A The logo color
B The flow's screen size
C Recipient and message content
D The file's creation date
Answer: C. Recipient and message content. Reviewing the recipient and content helps prevent misdirected or incorrect communications.
Which action usually needs human approval in a high-impact workflow?
A Formatting a temporary report
B Changing a consequential record
C Sorting a local table
D Renaming a draft
Answer: B. Changing a consequential record. Actions with significant consequences should have appropriate authorization or human review.
How should a workflow handle low-confidence AI extraction from a document?
A Treat it as verified
B Write it to every system
C Delete the source
D Route it for human review
Answer: D. Route it for human review. Low-confidence extraction should be reviewed before it triggers consequential downstream actions.
Why validate AI-generated fields before using them in a workflow?
A AI fields cannot be text
B Validation changes model weights
C Outputs may be wrong
D Prompts prevent every error
Answer: C. Outputs may be wrong. Validation checks expected structure and values before generated content is passed to another system.
What risk arises when untrusted document text is passed into an AI automation?
A It changes the timer
B It may contain manipulative instructions
C It removes all connectors
D It guarantees accurate extraction
Answer: B. It may contain manipulative instructions. External documents may contain prompt-injection instructions; systems should treat document content as untrusted data.
Which pattern is better for a stable, simple task?
A Unbounded autonomous actions
B A model with every permission
C An unclear prompt
D A deterministic rule
Answer: D. A deterministic rule. A fixed rule is often easier to predict for stable tasks with clear input conditions.
When might an AI step be useful in an automation?
A Interpreting varied unstructured text
B Adding two fixed numbers
C Renaming a known file
D Checking a static field
Answer: A. Interpreting varied unstructured text. AI can help interpret varied text, but the output needs validation before downstream use.
Why should each workflow step have a clear owner?
A Increase prompt complexity
B Support maintenance and incident response
C Change output format
D Remove audit logs
Answer: B. Support maintenance and incident response. An owner can review changes, respond to failures, and maintain connections or permissions.
How can workflow version history help an administrator?
A Prove every run correct
B Hide rejected actions
C Compare or restore versions
D Replace monitoring
Answer: C. Compare or restore versions. Version history supports change review and recovery, but it does not by itself prove the flow is correct.
What should be documented for a business-critical automation?
A Purpose, owner, and failure path
B Only its display name
C A private password
D The designer's theme
Answer: A. Purpose, owner, and failure path. Documentation should explain why the workflow exists, who maintains it, and what happens when it fails.
Why review a workflow after a connected service changes?
A The trigger always stops
B Permissions become irrelevant
C Past logs are deleted
D Fields and behavior may differ
Answer: D. Fields and behavior may differ. Connector updates may change available fields, authentication, or behavior, so existing steps should be checked.
Which approach helps detect a workflow that stopped running?
A Wait for a customer complaint
B Hide the flow status
C Monitor expected run activity
D Remove all notifications
Answer: C. Monitor expected run activity. Monitoring expected activity or heartbeats can show when an automation is no longer running.
What should an alert include when a workflow fails?
A Run identifier and failed step
B A user's secret key
C Unrelated personal data
D Only a decorative icon
Answer: A. Run identifier and failed step. A useful alert identifies the run and failure point while avoiding unnecessary sensitive data.
How can an automation handle a repeated webhook delivery safely?
A Process every copy
B Disable the endpoint
C Assume delivery is unique
D Check its event identifier
Answer: D. Check its event identifier. Webhook providers may retry delivery; deduplication by event identifier helps avoid repeated side effects.
Why should retry limits be explicit?
A Guarantee external availability
B Avoid endless processing loops
C Hide temporary failures
D Remove message history
Answer: B. Avoid endless processing loops. Defined attempt and time limits prevent a failing item from being retried indefinitely.
What does a workflow audit log help answer?
A Who changed what and when
B Which colors were preferred
C How long a prompt is
D The model's private reasoning
Answer: A. Who changed what and when. Audit logs help trace changes and actions for operational review.
When should an automation be disabled temporarily?
A When its name is long
B When a report is sorted
C When a test passes
D When actions repeat unsafely
Answer: D. When actions repeat unsafely. Pausing a faulty flow can prevent more unintended actions while the cause is investigated.
Which design choice makes an automation easier to maintain?
A One giant opaque action
B Hidden conditions
C Small, clearly named steps
D Unlabeled branches
Answer: C. Small, clearly named steps. Clear, modular steps make the workflow easier to review, troubleshoot, and update.
How should changes to a production workflow be introduced?
A Replace it without notice
B Roll out, monitor, and review
C Disable all logs
D Skip a rollback plan
Answer: B. Roll out, monitor, and review. A controlled rollout with monitoring and a recovery plan limits the impact of mistakes.
Why record assumptions in an automation's documentation?
A Increase the trigger count
B Prove output accuracy
C Clarify expected input and behavior
D Change service permissions
Answer: C. Clarify expected input and behavior. Documented assumptions help maintainers recognize when the workflow's conditions or source data have changed.
What should happen if a workflow cannot determine the correct path?
A Choose randomly
B Pause and request review
C Run every branch
D Pretend it completed
Answer: B. Pause and request review. Uncertain routing should be escalated or held for review rather than triggering arbitrary actions.