AI Data: Let Your Documents Tell You What You Need to Know
At its core, the Willow360 AI Data action helps you ask questions about a document and store the answers as structured data, also known as metadata. Once captured, this data can be reused throughout the workflow, including for conditional routing and decision-making.
Documents are classed as files such as .pdf .doc .docx .txt .odt and image representations of documents usually from document scans e.g. .tif .tiff .jpg .jpeg .png
What AI Data Does
AI Data analyses the content of a document using OCR combined with AI-based understanding. Instead of extracting fixed fields, you simply describe what you want to know, using natural language.
For each piece of information you want to capture, you define:
- A name, which is how the data will be referenced later.
- A question, which tells the AI what to look for.
- A data type, such as Text, Number, Date, Email, or Yes/No.
The goal is simple. Ask only for the information your workflow actually needs.
Real-World Examples
Example 1: Processing Invoices Automatically
You might ask the AI:
- "What is the invoice number?"
- "What is the total amount on this invoice?"
- "What is the supplier name?"
- "What is the invoice date?"
Once captured, this information can be used to:
- Rename the file, for example Invoice 45821 - £1,240.00
- Store the document in the correct folder
- Decide whether approval is required
Conditional processing example: If the Total Amount is greater than £1,000, send the document for approval. If it is £1,000 or less, skip approval and archive automatically.
Example 2: Checking Whether a Document Is Signed
For contracts or forms, you might only need a simple Yes or No answer:
- "Has this document been signed?"
- "Is there a handwritten signature visible?"
Conditional processing example: If Signed = No, send the document to a signing action. If Signed = Yes, continue straight to storage or notification.
This avoids unnecessary manual checks and ensures unsigned documents never slip through.
Example 3: Identifying Who the Document Belongs To
AI Data is also useful for identifying people or entities:
- "What is the full name on this passport?"
- "What is the customer email address shown?"
- "What company name appears on this document?"
This data can then be used to:
- Route documents to the correct team
- Automatically address emails
- Populate later forms or stamps
Example 4: Classifying Documents Automatically
Sometimes the most important question is simply what type of document this is:
- "Is this an invoice, receipt, or contract?"
- "What best describes this document?"
Conditional processing example: If Document Type = Invoice, follow the invoice workflow. If Document Type = Contract, send the document for legal review. If Document Type = Receipt, archive immediately.
This allows a single workflow to intelligently handle many document types.
Using Templates to Get Started Faster
To make things easier, Willow360 provides AI Data templates for common use cases such as invoices, receipts, ID documents, contracts, document classification, and smart document summaries.
Templates are a starting point, not a requirement. When you apply one, it simply pre-fills a set of common questions. You can remove anything you do not need, adjust questions to better match your documents, and add your own more specific queries.
The most effective AI Data configurations are usually focused and minimal, capturing only the data your workflow truly depends on.
Choosing Your AI Power Setting
AI Data offers two power settings: Regular and Pro.
Regular mode handles the majority of use cases well and costs 1 credit per file processed.
Pro mode uses a more powerful AI engine and is better suited to complex documents, dense layouts, or queries that require deeper reasoning. It costs 2 credits per file processed.
You can switch between the two settings during workflow configuration. The total workflow cost updates live as you make changes, so you can see the impact before saving.
Turning AI Data into Workflow Decisions
Once the AI Data action finishes, all requested information is guaranteed to be available, either read automatically or confirmed by a user if something was missing.
From that point onward, the captured data behaves like any other metadata in Willow360:
- It can be used in conditions.
- It can appear in file names.
- It can be inserted into emails, stamps, or storage paths.
- It can determine who needs to act next.
- It can be recorded into a structured report using the Export Data to CSV action.
This is what makes AI Data powerful. It does not just read documents. It enables smarter, self-directing workflows based on document content.
A Simple Rule of Thumb
When configuring AI Data, always ask yourself:
"What information do I need from this document to make my workflow work?"
If the answer is clear, AI Data will do the rest.
Related Articles