Anthropic · Technologies
Claude Compliance API
Answer the question your regulator will eventually ask.
The Compliance API gives enterprise organisations real-time programmatic access to Claude usage data and customer content. If you work somewhere that has to evidence how AI is being used, this is the difference between believing your controls work and being able to demonstrate it.
COMPLIANCE FEED · EXAMPLE ORGANISATION
CHECK
SOURCE
FREQUENCY
STATUS
Usage by team
Compliance API
Real time
ACTIVE
Retention rules
Compliance API
Daily
ACTIVE
Flagged content
Compliance API
Real time
ACTIVE
Access reviews
Directory + API
Monthly
DUE
Third-party tools
Manual today
Ad hoc
GAP
Sample organisation
active · due · gap
In plain terms
Everyone says AI use is governed. Few can show it.
The difference matters the first time somebody formally asks.
Without it
- Usage is described in a policy nobody measures
- Evidence is gathered by hand before an audit
- No way to see AI use as it happens
- Incidents are found late, or by someone else
With it
- Usage data flows into your own systems
- Evidence is produced from a live feed
- Patterns are visible while they can be acted on
- Issues surface early enough to matter
What CG TECH can do with the Compliance API
For businesses that have to prove it, not just say it.
Regulated businesses, government suppliers and anyone whose customers audit them. If nobody has ever asked you to evidence AI use, this is probably not your first priority.
A feed is only worth having if it lands where your team already looks. We connect it into your reporting or security tooling rather than creating another screen nobody opens.
It covers Claude. Staff using other AI tools, approved or not, sit outside it. Microsoft Purview covers that ground, and most businesses need both to answer the question honestly.
Raw data is not an audit answer. We agree what needs reporting, how often, and what a reviewer should be shown, so the work is done before the request arrives.
How an engagement runs
From a policy document to evidence you can hand over.
01
Agree the questions
We work out what you would actually be asked to prove.
02
Connect the feed
Usage data goes into your reporting or security tooling.
03
Build the reporting
We turn it into something a reviewer can read.
04
Set the rhythm
Who looks at it, how often, and what triggers action.
Questions we hear a lot
Common questions about the Compliance API
What is the Claude Compliance API?
It gives enterprise organisations real-time programmatic access to Claude usage data and customer content, so AI use can be monitored and evidenced through your own systems rather than described in a policy.
Who actually needs it?
Businesses with a regulatory obligation, government suppliers, and anyone whose customers audit their controls. If nobody has ever asked you to evidence AI use, there are probably higher priorities first.
Does it cover AI tools other than Claude?
No. It covers Claude. Staff using other tools sit outside it, which is why most businesses pair it with Microsoft Purview to see AI use more broadly. We would rather say that than let you assume one product covers everything.
Which plan is it available on?
It is an enterprise capability, so it comes with Claude for Enterprise rather than the individual plans. Claude for Enterprise is also available to organisations operating under HIPAA.
Is this the same as monitoring our staff?
It is worth being clear about the difference, and being clear with staff. The purpose is evidencing that controls work and catching problems early. How far it goes is a decision for you, and it should be written down and communicated.
What do we do with the data?
Feed it somewhere your team already looks, usually existing reporting or security tooling. A separate dashboard nobody opens is a common and expensive mistake.
How long does it take to set up?
Connecting the feed is quick. Agreeing what needs to be evidenced, and building reporting a reviewer would accept, is the part worth allowing time for.
Can you help with the wider AI governance work?
Yes. This is one input. The rest is deciding what staff may use, what is off limits, who approves new tools and how that gets reviewed. We do that work alongside the technical setup.
Ready when you are
Could you evidence how AI is used across your business today?
Most organisations cannot, and find out at the worst moment. A discovery session covers what you would be asked, what you can show now, and what the gap would take to close.
What to expect
- A consultant replies within 4 business hours
- Session booked to understand your requirements
- We will provide you with a fixed price quote