Do you copy our records out of our network?

No. Discovery builds an inventory of systems and fields. The records themselves stay where they already are.

What does a discovery run produce?

A live list of where personal data sits, how it is classified, and whether a rights request can look it up exactly.

Platform

An inventory of systems and fields, built from the catalogs you already run. Records stay in your network. Further connectors can be added.

Inside your network

Inventory

customers.email

Contact identifier. Indexed for rights. Rows stay put.

First find the stores. Then name the personal data. Then mark what a rights request can look up exactly. A guess is not filed as found.

Step 1
Locate
Looks at: The catalogs and stores you already run
Produces: An inventory of systems and fields. Not a copy of the rows.
Step 2
Classify
Looks at: Which fields hold personal data
Produces: Labels the fiduciary can act on: contact, financial, child.
Step 3
Confirm
Looks at: Whether a rights lookup can hit an exact key
Produces: A list ready for access and erasure. Guesses are not filed as found.

The control plane holds names of systems and fields. It does not take a copy of the rows.

Your network

Databases, warehouses, and files stay on your side. Row values are not shipped out to build the inventory.

Control plane

Names of systems and fields, classification, and the work to run next. Not a copy of the records.

Live types on the moving rail. Records stay in your network.

Further stores use the same inventory. They are not claimed as live until they are connected on an estate.

Reasonable security. Safeguards start with knowing which fields hold personal data and who can reach them.
Rights. Access and erasure use exact lookups. A sampled search that misses a row is not a completed request.

No. Discovery builds an inventory of systems and fields. The records themselves stay where they already are.
A live list of where personal data sits, how it is classified, and whether a rights request can look it up exactly. It is not a spreadsheet of system names someone typed last year.
Relational databases, warehouses, catalogs, NoSQL, object storage, streams, major SaaS, cloud inventory, logs, and backup vaults. Further systems can be added without changing how the inventory is read.