Skip to content
Demiton

Create Your First Workflow

This guide walks through the process of creating a simple workflow that retrieves labour entries from a field system, transforms them, and posts the aligned records to a financial system.

Step 1: Define the workflow

Workflows are ordered sequences of steps. Each step performs one operation using a workflow verb.

Example structure:

1. FETCH labour entries from Assignar
2. TRANSFORM records into LabourEntry business objects
3. PUSH aligned records to Business Central

This structure ensures data flows through a predictable, observable pipeline.

Step 2: Retrieve data from a system

The first step retrieves data from an external system using the FETCH verb.

FETCH → assignar.site_diaries

The Assignar adapter handles authentication, API communication, and response normalization. The result is wrapped in a Sovereign Envelope and stored in the pipeline context.

Step 3: Transform the data

External system records often need normalization before they can be used elsewhere. A TRANSFORM step converts raw records into canonical business objects.

TRANSFORM → extract_labour_entries

This step converts raw site diary records into structured LabourEntry business objects with fields including:

  • worker
  • project
  • date
  • hours
  • cost category

Step 4: Apply governance checks

Before data is written to downstream systems, governance rules validate it.

GOVERN → validate_project_mapping

Common checks include:

  • validating project identifiers exist in the ERP
  • verifying worker mappings
  • enforcing cost category rules

Invalid records are stored in an exception dataset rather than propagating to downstream systems.

Step 5: Write records to the target system

Once records are validated and transformed, write them to the target system using PUSH.

PUSH → business_central.project_ledger

The Business Central adapter handles authentication, entity mapping, and transaction handling. All PUSH operations support simulation mode — run the workflow with dry_run: true first to validate behavior before writing to production.

Step 6: Observe execution

Every workflow run produces a WorkflowRun record with full execution metadata:

  • run identifier
  • step execution results
  • execution time per step
  • system responses

You can view this under Workflows → Runs in the platform.

Summary

A typical workflow follows this pattern:

VerbPurpose
FETCHRetrieve records from an external system
TRANSFORMConvert records into canonical business objects
GOVERNValidate and enforce policy rules
PUSHWrite aligned data into downstream systems

Next steps