You are not looking at the latest version of the documentation. Check it out there.
You are looking at draft pre-release documentation for the next release of Flows for APEX. Product features documented here may not be released or may have different behavior to that documented here. You should not make any purchase decisions based on this draft documentation..

Using Manual Adhoc Sub Processes

Manual Adhoc Sub Processes are designed for knowledge-worker-driven execution.

Editions: Manual Adhoc Sub Processes are available only in the Flows for APEX Enterprise Edition.

The subprocess does not prescribe one fixed internal sequence. Instead, it presents the user with the activities that are currently startable, and the user decides which activity to perform next.

The AHSP continues until its completion condition is met.

The Manual AHSP Pattern

In a manual AHSP:

  • the model defines the possible activities
  • the runtime determines which activities are currently startable
  • the knowledge worker selects the next activity
  • the subprocess updates its current situation after each completed activity
  • the subprocess finishes when the completion condition is satisfied

This pattern is useful when expert judgment is required to decide what should happen next.

Typical User Journey

The normal manual AHSP flow looks like this:

  1. the AHSP becomes active
  2. the user sees the currently startable activities
  3. the user chooses one activity to perform
  4. the runtime collects any required input parameters
  5. the activity executes and returns outputs and summary information
  6. the current situation is updated
  7. the completion condition is checked
  8. if the AHSP is not finished, the next set of currently startable activities is shown

This cycle repeats until the completion condition becomes true.

Designing for Knowledge Workers

Manual AHSPs work best when users can quickly understand:

  • what activities are available now
  • why an activity is relevant
  • what information each activity needs
  • what happened as a result of completed activities
  • how close the subprocess is to completion

That means the model should provide more than just technical activity names.

Good manual AHSP design includes:

  • clear activity names
  • concise activity descriptions
  • sensible display ordering or grouping
  • strong input parameter definitions
  • good completed-activity summaries

Input Parameters and User Input

Manual AHSP activities often need user-provided input.

Task Parameters are therefore central to manual AHSP design.

  • input parameters define what the activity needs from the user
  • layout metadata can improve how those inputs are rendered
  • output parameters define what the activity returns to the subprocess

See Task Parameters and Input Parameter Layout Guide.

Reading the Current Situation

After each activity completes, the manual user should be able to understand the updated situation without having to inspect raw logs or internal state.

In practical terms, this means the AHSP should present:

  • what has already been done
  • what the important results were
  • what remains uncertain or unresolved
  • what can be started next

The better the completed activity summaries are, the easier it is for users to make good decisions.

Completion in Manual AHSPs

Even in a manual AHSP, completion should not depend on the user guessing when to stop.

The model should define a real completion condition so the subprocess finishes when the intended outcome has been reached.

This gives the knowledge worker flexibility in activity order, while still keeping the subprocess controlled and auditable.

Good Modeling Practice

For manual AHSPs, good practice is to:

  • start with a small, meaningful initial activity set
  • avoid large numbers of poorly differentiated activities
  • make start conditions explicit and understandable
  • use repeatability only when it reflects a real business need
  • make the completion condition specific and testable
  • design each activity to return a useful summary of what changed

Relationship to AI-Driven AHSPs

Manual AHSPs and AI-driven AHSPs share the same base model.

If your AHSP is designed well for manual use, it is usually much easier to evolve it later into an AI-assisted or autonomous pattern.

See AI-Driven Adhoc Sub Processes.

Updated: