Cresec

By role

Approving AI tools your colleagues built: a guide for IT and security

For IT and security, homemade AI tools are a familiar problem in a new form: useful software, built outside IT, holding someone's login. Blocking it rarely works — people find a way around. Approving it is easier when you can see who uses it, limit what it can do, and switch it off per person.

Common workflows

Approve a team tool without shared logins

People sign in through your existing company login (like Okta or Microsoft Entra). Each person connects their own apps; the tool never stores a password or key.

Apps it uses: Your company login, The apps the tool uses

Why it's hard to share today: Today the usual choice is a shared account or an outright ban.

One record per action

Every action is written down once, naming the person and the tool version, linked to the record before it and checkable offline.

Apps it uses: Your logging or security tools

Why it's hard to share today: Homemade tools usually leave no trail at all.

You bring the logic. Cresec does the rest.

Common questions

What if Cresec is unavailable?

Cresec is designed so tools fall back to read-only rather than break: people can keep reading and drafting, and nothing is sent or changed without a record.

How hard is it to stop using Cresec?

Leaving is designed to be simple: it's a small configuration change in the tool, and you keep your records and can still check them.

Which company logins does Cresec work with?

Your existing identity provider, such as Okta or Microsoft Entra. People sign in the way they already do.

Do we host the tools?

The tools keep running wherever their builders put them. Cresec sits between them and your apps, so you control access without hosting anyone's code.

Can we see every tool in use?

Yes. Every tool, version, person and action shows up in one place, so shadow tools become visible — and approvable.