KB-2421 AI Skill Extraction Training Fails With 'There Were Not Enough Labeled Fields to Train a Model' Error for Specific PDF Form Fields

Symptoms

When attempting to extract a field from a fillable PDF document, the AI Skill training fails with the error:

There were not enough labeled fields to train a model. Label each field in at least 5 documents.

Removing the problematic fields allows the model to train successfully.

Cause

The source documents are fillable PDF forms, and field values exist only as form data.

Action

1. Manually flatten the PDFs (e.g., printing to PDF) used to train the model.

2. Configure a process model to automatically flatten all future input PDFs using the Fill PDF node from the PDF Tools plugin before they are processed by the AI Skill.

Affected Versions

This article applies to all versions of Appian.

Last Reviewed: October 2026