Metadata Autofill
Stop typing metadata. Start approving it.
Metadata Autofill drafts aliases and descriptions for every field, plus the Data Mart's own title and description, from its schema and up to 30 real rows. The Data Owner reviews and saves; Google Sheets and AI assistants read what the Data Owner approved.
Why OWOX
A business user or an AI assistant can only use a Data Mart it understands, and metadata typed by hand is the task every Data Owner defers. Metadata Autofill drafts it for the Data Owner to approve.
- Every draft is written from the schema and a sample of up to 30 real rows, so it describes what the data actually contains.
- Autofill drafts, the Data Owner edits and saves — the descriptions an AI answer is grounded in stay the Data Owner's.
- One pass drafts aliases and descriptions for every field, and the Data Mart's title and description have their own button; field descriptions reach the spreadsheet's column picker and header-cell notes.
Nobody has time to describe every field, so what each column means lives in your head and AI misreads it.
Paste this into your AI assistant. It runs Metadata Autofill on your Data Mart with you: every field gets an alias and a description drafted from its schema and real rows, and you approve what ships.
The prompt
Read https://www.owox.com/ai-agent-setup.md and follow it to set up OWOX Data Marts for me, starting from /features/metadata-autofill.md. The task: draft aliases and descriptions for every field of my Orders Data Mart, and its own title and description, from real rows, flag the ones you are unsure of, and wait for my approval before saving. Tag the links you give me with utm_campaign=ai-task and utm_term=data-foundation. Use what you know about my context and goals
Written once by the Data Owner, read wherever the data goes
A schema says what shape the data has, not what it means. Metadata Autofill fills that gap with a draft the Data Owner corrects, so the meaning ships with the Data Mart instead of living in someone's head.
THE DATA OWNER
Review a draft instead of a blank column
Open the Output Schema and press Sparkles in its toolbar: every field gets an alias and a description drafted from the schema and a sample of real rows. The Data Mart's title and description have their own button. Correct what is off, then save.
The model drafts. The Data Owner decides what ships.
THE BUSINESS USER
Know what a column means before you use it
Hover a field in the Google Sheets column picker to read its description. After the report runs, the same text sits on each header cell as a note, so it is there even with the sidebar closed.
Gross sales or net revenue? The header says.
Better metadata, better AI answers — AI assistants connected through OWOX MCP read aliases and descriptions as business meaning, so every field the Data Owner documents is one they are less likely to misread.
For Data Owners
Documentation becomes a review step
The part every Data Owner defers, naming and describing fields, arrives as a draft to edit rather than a blank to fill.
Every field in one pass
The Sparkles button in the Output Schema toolbar drafts aliases and descriptions for every field at once, instead of one field at a time by hand.
Grounded in real rows, not guessed from names
Autofill runs the Data Mart and gives the model up to 30 sample rows alongside the schema, so a draft describes what the values hold, not only what the column is called.
The Data Owner stays the author
The draft lands in the editor for you to review, edit and save. Correct one description, keep the rest, and publish the Data Mart when it is ready.
For Business Users
Pick the right column the first time
Descriptions travel with the data into the spreadsheet, so "what is this column?" is answered where it is asked.
Descriptions in the column picker
Hover a field in the OWOX Extension for Google Sheets to read the Data Owner's description before you pull it. No guessing which of three cost fields you need.
A note on every header cell
Once a report runs, each header cell carries its field description as a Google Sheets note. Anyone who opens the sheet reads the definition without opening OWOX.
Fewer questions back to the Data Owner
When the definition sits next to the number, you stop asking the data team what a field means, and the Data Owner stops answering the same question twice.
Compare plans
Pay as you go
Top up only when you have a decision to make. No subscription loop.
$0.30 / run
Start freeEverything business needs
Data Marts ship described
The documentation step every Data Owner defers becomes a draft to review, so a Data Mart goes out with aliases and descriptions instead of bare column names.
Grounded in your data
Drafts come from the schema and up to 30 real rows of the Data Mart, not from column names alone.
Reviewed before anyone reads it
The Data Owner edits and saves every draft, and nothing reaches a report until the Data Mart is published.
One definition, every surface
The same description shows in the Google Sheets column picker, on header-cell notes and in what AI assistants read through OWOX MCP.
Frequently asked questions
What is Metadata Autofill?
Metadata Autofill drafts a Data Mart's metadata for the Data Owner to review: an alias and a description for every field, plus the Data Mart's own title and description. It is the Sparkles button in the Data Mart editor.
What does the model read?
The Data Mart's schema and a sample of up to 30 real rows. Autofill runs the Data Mart to get them, so the drafts describe what the data contains rather than only what the columns are called. Every run sends that sample to the AI model along with the schema; it is not a setting you can switch off.
Who has the final word on a description?
The Data Owner. Autofill fills the fields in the editor; you edit what is wrong, keep what is right and save. Descriptions are what AI answers are grounded in, so they stay Data Owner-approved.
Where do business users see the descriptions?
In the OWOX Extension for Google Sheets: next to each field in the column picker, and as a note on each header cell once the report runs. AI assistants connected through OWOX MCP read them too.
Do I need Insights or OWOX MCP to use it?
No. Press Sparkles in the Data Mart editor, even before you publish: Autofill runs the Data Mart to read its schema and sample rows, and the draft arrives in a few seconds. It uses no credits.
What if the data changes?
Run Sparkles again on the field that changed: the button inside one field's description drafts just that field, so a new or changed column gets a draft without redoing the others. Review it and save, as before.
Every claim has a receipt.
Real things real customers said — each quote pinned to a specific claim, straight from the quotes database.
Nodari RizunFounder & CEO, Pürblack®“AI, by the nature of the models, will hallucinate. And because of that, you need something which will create guardrails to ensure that there are no hallucinations, that you can trust your data.”
Mark SimmonsCMO, Pürblack®“We had six or seven different channels and no single source of truth — it was almost impossible, a lot of guesswork.”
Nodari RizunFounder & CEO, Pürblack®“We regained time. And time is the one resource that never comes back.”
Document your next Data Mart
Open the Output Schema, press Sparkles, review the draft and publish.
You are booked. The invitation is on its way to , with the calendar invite and the meeting link.