Skip to content
DMDQP AI Capability Hub

Generate a Collibra Asset Description

Generate a Collibra Asset Description

Context

When onboarding a data asset into Collibra — whether a dataset, report, data product, or business term — you need a concise, non-technical description that tells data consumers what the asset contains, who uses it, and what business decisions it supports. Writing these by hand at scale is time-consuming and inconsistent. This prompt produces a business-audience description you can paste directly into Collibra (with a data steward review before publishing).

Use this prompt with build-cli when you have a schema file or data dictionary to provide as additional context. Use Gemini in the browser for one-off descriptions.

The Prompt

Write a Collibra asset description for the following data asset.


The description should:
- Be written for a business audience (not technical jargon)
- Explain what the asset contains, who uses it, and what business decisions it supports
- Be 2-4 sentences
- Not include implementation details (database name, table name, column types)


Asset name: [ASSET NAME]
Asset type: [Dataset / Report / Data Product / Term / ...]
What it contains: [BRIEF TECHNICAL DESCRIPTION — this is your input, not the output]
Primary consumers: [WHO USES THIS]
Business purpose: [WHAT DECISIONS OR PROCESSES DOES IT SUPPORT]

Usage Instructions

  1. Fill in all five placeholders ([ASSET NAME], [Asset type], [BRIEF TECHNICAL DESCRIPTION], [WHO USES THIS], [WHAT DECISIONS OR PROCESSES DOES IT SUPPORT]) with the real values for your data asset.
  2. Paste the completed prompt into build-cli or Gemini.
  3. Review the generated description for accuracy — AI outputs can misrepresent business context.
  4. Have a data steward review and approve the description before publishing it in Collibra.

Batch mode: To generate descriptions for multiple assets at once, list the assets in a table in your prompt (one row per asset, columns for name, type, what it contains, consumers, and business purpose) and ask for a description per row.

Example Output

Given inputs: Asset name = “Customer Master Dataset”, Asset type = “Dataset”, What it contains = “Deduplicated customer records from CRM, ERP, and MDM systems”, Primary consumers = “Sales, Finance, and Data Governance teams”, Business purpose = “Supports customer 360 reporting, sales territory planning, and regulatory compliance”.

Expected output (paraphrased):

The Customer Master Dataset provides a single, deduplicated view of all customer records consolidated from our CRM, ERP, and MDM platforms. It is used by Sales, Finance, and Data Governance teams to power customer 360 reporting, support sales territory planning, and meet regulatory compliance requirements. Always verify record counts with the MDM team before using this dataset in external-facing reports.