Example Dooap Actions

These examples show practical starting points.

Use them as templates you can adapt, not as fixed blueprints.

Invoice Data Extraction

• Input: invoice PDF.
• Steps: parse document, extract fields, validate totals.
• Output: structured JSON for downstream processing.

Why this is useful:

• Reduces manual data entry.
• Produces consistent output shape for integrations.

Vendor Matching

• Input: vendor name and account details.
• Steps: normalize values, call vendor search tool, resolve best match.
• Output: matched vendor identifier and confidence.

Why this is useful:

• Improves match quality for noisy source data.
• Makes uncertain matches visible through confidence signals.

Approval Routing

• Input: invoice amount and business context.
• Steps: apply policy logic, choose approver group, send notification.
• Output: routing decision with audit metadata.

Why this is useful:

• Turns approval rules into transparent, repeatable Action Flow behavior.
• Creates a reliable audit trail for compliance and support.

Suggested Starter Variations

If you want to expand from the examples above:

• Add a low-confidence branch that routes to manual review.
• Extract metadata fields for easier run filtering.
• Trigger a downstream Action for specialized post-processing.

How to Use Examples

• Start from the closest example.
• Keep your first version simple.
• Validate with real-world sample inputs.

Common Adaptation Mistakes

• Copying prompts without adjusting domain terms.
• Skipping validation for required business fields.
• Combining too many responsibilities into one Action.

Treat examples as scaffolding, then refine for your exact process and data quality.

---