Know what every column means, and how sensitive it is.
Plain-English descriptions and sensitivity ratings for every column, plus quality rules where your business needs them. Nothing is final until a person approves it.
Your database
Never changed
Metadata Enrichment Engineinstalled in your Azure subscription
Profile
Fill, values and ranges for every column, year by year
Classify
Write rules
Quality rules for the columns your requirements name, as SQL
Test
Every rule checked year by year; every problem confirmed on all of your data
Describe
A plain-English description of each column, from its profile and your glossary.
Review and approve
Nothing is final until a person approves it.
- Business descriptionPlain English, approved by a person
- Sensitivity classificationSet by stated rules, never by AI
- Quality rulesWhere your requirements need them, with SQL and evidence
You know what data you have. Not what it means.
Column names like cust_dt_ltv_fl tell no one anything. Nobody is sure which columns hold personal data, and quality is judged by feel. So every report, audit and AI project starts by digging it all up again.
Three answers your team can use
A plain-English description of every column, approved by your data owner.
Remove or mask, handle with care, or use as is. Set by stated rules, not guesswork.
Quality rules where your business needs them, with proof for every problem found.
What your team works with
1We measure
Every column profiled: how full it is, its formats, ranges and duplicates, year by year where the table has a date.
2We propose
Rules for the columns you said matter, each with the evidence behind it.
3You decide
Keep, change or reject every rule. Nothing becomes a rule without you.
"Which numbers do you double-check before you trust a report?"
Contribution dates. Anything outside the two-year election cycle needs a second look.
Contribution date must fall inside the filing's two-year election cycle
indiv20.transaction_dtValidityTimelinessEvidence: 0.4% of rows fall outside| Column | Business description | Sensitivity | Handling | Quality rules |
|---|---|---|---|---|
cmte_id | The committee that received the contribution, as the FEC's committee ID. | Internal | Use as is | Must match a known committee |
name | The contributor's name as filed, last name first. | Confidential | Handle with care | None required |
zip_code | The contributor's ZIP code, five or nine digits. | Confidential | Handle with care | Five or nine digits |
employer | The contributor's employer, as reported by the contributor. | Confidential | Handle with care | None required |
occupation | The contributor's occupation, as reported by the contributor. | Confidential | Handle with care | None required |
transaction_dt | The date the contribution was received. | Internal | Use as is | Inside the election cycle Never in the future |
transaction_amt | The amount contributed, in US dollars. Negative for refunds. | Internal | Use as is | Not empty |
sub_id | The FEC's unique ID for this record. | Internal | Use as is | Unique |
indiv20.memo_textBased on: your documentationDrafted from the column's profile and the FEC data dictionary.
A note the filer added to the contribution, such as why it was refunded or whom it was earmarked for. Free text; may name people.
Public is set only by a person, never by the engine.
| Column | Level | Identifies a person | Why |
|---|---|---|---|
name | Confidential | Directly | A person's name |
zip_code | Confidential | Indirectly | Where a person lives |
occupation | Confidential | Indirectly | Describes a donor |
memo_text | Confidential | Possibly | Free text that may name people |
cmte_id | Internal | No | A committee, not a person |
Contribution date must fall inside the filing's two-year election cycle
Public FEC campaign finance data. Screens are illustrative.
One fixed fee. No subscription.
One data domain · six weeks
- Installed in your Azure, read-only on Snowflake or Azure SQL
- A description and a sensitivity rating for every column
- Quality rules for the columns you need
- Your team trained to run it again
No license fee. The engine stays installed and you can keep running it. Each run uses your own Azure and database compute.
Add-ons
Most rules carry over, so it costs less than the first.
Everything approved, loaded into your data catalog.
Databricks and Redshift are next, with no date set yet.
What leaders ask us first
We already have a data catalog.
Will it change anything in our database?
Does AI decide what is sensitive?
Why doesn’t every column get a quality rule?
What happens after the six weeks?
We are on Databricks or Redshift.
Start with one data domain
Six weeks, one fixed fee, and the engine stays with you.