---
name: tbt-ai-data-dictionary
description: "Explain field meaning, unit, key and missing-value conventions before analysis for Give AI a data dictionary before it combines your files. Use permitted source material; this kit does not authorise external actions."
---

# Give AI a data dictionary before it combines your files

Explain field meaning, unit, key and missing-value conventions before analysis.

Use the supplied source as authority. Check the account supports the actual file or image operation before attempting it.

## Task instruction

Before combining these supplied files, use this data dictionary:
[field meaning, type, unit, grain, key and missing convention].
Confirm the intended join and scope. Preserve text identifiers including
leading zeros, valid numeric zero and explicit missing values.
Check required key uniqueness and unmatched records. Do not infer units,
aggregate grain or duplicate policy from column names. Ask a focused
question for any definition needed before the calculation can be valid.

## Source boundaries

- Customer IDs are text with leading zeros
- Amounts are pence
- Open rows only; unique customer keys

## Review and correction

001 is not automatically 1. Duplicate lookup keys or unmatched IDs need an explicit rule before totals.

The result changed [identifier/unit/missing state/grain] from the supplied
definition. Restore that definition and redo only the affected join or
conversion. Preserve source rows and the agreed filter. Leave ambiguous
duplicate matches unresolved rather than pick one. Show the affected
record IDs and revised counts; do not hide the problem in a zero or total.

Use references/practice-source.json for the labelled example. The practice HTML is a local teaching tool; it does not call an AI service. Apply this method to the user's supplied facts rather than copying fictional example values. Do not infer approval or permission to publish.