Founding offer · 5 spots The first 5 customers get 3 months free — a personal setup and priority on the features you ask for. See prices
Data security governance for Snowflake, dbt and Airflow

Turn the lights on in your data.

Know which of your Snowflake data is sensitive, who can read it, and whether that access looks too broad — with a named owner for every risk. More than a data catalog. Reads metadata only.

Built around the data security practices in DAMA-DMBOK.

Reads metadata from Snowflake dbt Airflow
Your data, mapped
fine needs attention exposed unknown
Sample data
Before and after

Most teams run their data in the dark

You have data, and you have opinions about it. What you don't have is a map.

In the dark

"We have customer data… somewhere."

In the light

Every column has a security level — and the sensitive ones are on one screen.

Assumption

"Only a few people can see payroll."

Evidence

6 roles and 1 named user can read payroll. Your policy allows 3.

Random facts

"Finance owns that, I think." "Masking is set up, mostly."

A strategy

Every sensitive table has an owner, a level and a next step — ranked by risk.

Overconfidence

"We're too small for anyone to ask."

Awareness

Customers, auditors and regulators already ask. Now you can answer.

Silence

"No one said it was a problem."

Accountability

Anna accepted this risk on 2 March. It comes back to her on 2 June.

Sample data

How it works

From the dark to a plan, in four steps

Step 1 · Dark

You have data and assumptions

Tables in Snowflake, opinions in people's heads, a spreadsheet nobody trusts.

Step 2 · Light

Every column gets a level

Public, Internal, Confidential or Restricted — suggested for you, decided by you.

Step 3 · Map

You see who can read it

Roles and users with access to each sensitive table, and what looks too broad.

Step 4 · Strategy

Owners decide, risks get dates

Sign off, accept with a reason and an expiry, or delegate. A plan you can show.

Three jobs

Define. See. Act.

The data-layer side of data security, for teams without a security department.

1 Define

Every column gets a security level

AB Data Management suggests a level for every column from its name, description and neighbouring columns. Your team reviews it — and your decision always wins.

  • Four levels: Public, Internal, Confidential, Restricted.
  • Spots combinations that can identify a person, not just columns called "email".
  • Tags such as PII, PCI, PHI and Regulated, plus your own.
  • Lowering a level needs a written reason.

Try a column name

Live demo

Same name-pattern rules as the product. Suggestions only — your decision always wins. Nothing you type leaves this page.

2 See

Who can read it — and whether that looks too broad

We read your Snowflake grants and show every role and user that can reach each sensitive table — directly, or through its database, schema or a shared database.

  • A flag when more roles can read it than your policy allows (by default 3 for Restricted, 10 for Confidential).
  • A flag when it's open to everyone, or granted to one person directly.
  • Sensitive columns with no masking policy. If we can't check, we say Unknown.
  • You make the change in Snowflake. We show you where.

Who can read payroll.salaries?

payroll.salariesRestricted Finance analysts HR admins Payroll service Marketing Sales ops Everyone (PUBLIC) Direct: j.smith
Looks too broad · 6 roles + 1 user · policy allows 3 Within policy · 3 roles · signed off by Anna K. Sample data
3 Act

An owner and a date on every risk

Each owner sees only what they own, what's risky, and what to do — no Snowflake access needed.

  • Sign off, accept a risk (reason and expiry required), or delegate it.
  • When an accepted risk expires, it comes back on its own.
  • Recertification campaigns ask owners to re-confirm.
  • Owner decisions are recorded. Export your posture summary as CSV or HTML.

payroll.salaries · accepted risk

Sample data
AK
Anna K.
Accepted risk

"HR roles are being migrated — access will be trimmed by June."

Accepted 2 March · expires 2 June
Expired — back on Anna's list

The exception ran out, so the risk is open again and Anna is asked to decide.

Nobody has to remember. Risk accepted forever is just risk forgotten.
Not just a catalog

A catalog tells you what you have. We tell you what's exposed.

Drag the handle to compare the same table.

A data catalog tells you

payroll.salaries

Monthly salary per employee. Loaded nightly from the HR system.

Owner
HR team
Security level
Not set
Who can read it
Not shown
Masking
Not shown
Risk
Not shown
Decision on record
None
AB Data Management also tells you

payroll.salaries

Monthly salary per employee. Loaded nightly from the HR system.

Owner
Anna K. · People domain
Security level
Restricted · PII
Who can read it
6 roles + 1 user · policy allows 3
Masking
2 sensitive columns without a policy
Risk
High — access open to everyone
Decision on record
Accepted by Anna · back on 2 June
← Catalog only · AB Data Management → Sample data
Honest by design

Unknown, never a false all clear

Unknown means unknown

If we can't verify masking or access, we say so. We never mark something protected without evidence.

Metadata only

We read names, descriptions, grants and policy names — never the rows in your tables. Refreshed every time you sync.

You stay in control

We don't change anything in Snowflake. We show you where the gaps are and who owns them; your team makes the fix.

What it does not do

×Change grants, masking or encryption in Snowflake
×Read the contents of your tables
×Certify compliance or replace an audit
×Replace network, device or endpoint security
Common questions

You might be thinking…

"Isn't this just another data catalog?"

The catalog is the foundation, not the point. What you get is a security view: where sensitive data is, what's unmasked, who can read it, and who owns each risk.

"Snowflake already has roles and masking policies."

Snowflake gives you the controls. It doesn't tell you which sensitive columns are missing them, or which tables more roles can read than they should. We show those gaps in plain language.

"We're too small for this."

Small teams get the same questions from customers, auditors and regulators as big ones — just without a security department to answer them. That's exactly who this is for.

"How do I know the findings are right?"

Every risk comes with a plain-language reason, levels are suggestions your team confirms, and anything we can't verify is marked Unknown.

The first 5 customers get 3 months free, a personal setup, and priority on the features they ask for.

Claim a founding spot
Lights on

Ready to see your data in the light?

We'll connect read-only to your Snowflake and show you which data is sensitive, who can read it, and what looks too broad.