Workbench / Issues

Wrong Field Types

The AI returns fields with values whose types do not match the expected schema, such as strings where numbers, booleans, arrays, objects, or enums are required.

What This Looks Like

The AI returns structured output where one or more fields are present, but the values use the wrong type. A field expected to be a number may be returned as a string, an array may be returned as a sentence, a boolean may be returned as “yes,” or an enum may be replaced with a similar freeform label.

Why It Matters

Wrong field types can pass casual human review but fail machine validation. They create problems at the boundary between natural language and structured systems, especially when the next step expects strict types for parsing, storage, automation, comparison, or execution.

Structural Signal

The field exists, but its value does not conform to the declared type contract. The issue is not missing information and not necessarily invalid JSON syntax; it is a mismatch between the expected field type and the generated value.

Common Triggers

When to Use This Issue

Use this Issue when fields are present but their values do not match the expected type required by the schema, validator, parser, importer, or downstream consumer.

When Not to Use This Issue

Do not use this Issue when required fields are missing, extra fields are added, or the entire response cannot be parsed. Use this Issue when the structure is present but the field types are wrong.

Category

Output

Primary Pattern

PAT-0160 — Schema Breakage

Declared Patterns

Derived Primary Lenses

Derived Secondary Lenses

Search Intents

Ontology Metadata

Code
ISS-0024
Version
ISS-0024@0.1.0
Ontology release
0.1.0
Updated
2026-05-10T00:00:00Z

History

No public history entries recorded.

View full ontology changelog