How to govern Claude.ai
Anthropic's assistant for drafting, analysis and long documents, and what to watch as its capabilities grow.
Anthropic · Reviewed monthly, and when Anthropic announces a capability change · Last reviewed 16 Oct 2026
Governance change Reviewed the app release notes from 3 September to 7 October. Anthropic has merged Cowork into every conversation: what was only possible in Cowork, including working in local files, browsing, computer use through the desktop app and scheduled tasks, is now available from any chat, with an Auto or Manual permission setting per conversation. Pro and Max first; Enterprise administrators get 30 days notice. Claude Cowork no longer has its own page here. Full changelog ↓How Claude.ai is used
We see seven common ways businesses use Claude.ai, and we govern each one on its own terms. If your team uses it differently, that is fine: you describe the use in your own words and Certrak works out the governance from what it actually does.
General productivity assistance
Consequence to people: MediumDrafting, summarising, brainstorming, internal Q&A - day-to-day help with internal work
Document analysis and review
Consequence to people: MediumReviewing contracts, reports, or technical documents
Marketing content creation
Consequence to people: LowDrafting marketing copy, social posts, ads, blog content
Customer communications drafting
Consequence to people: MediumDrafting customer emails, support replies, or other outbound communications - reviewed before sending
Code generation and review
Consequence to people: LowWriting or reviewing code, with a developer in the loop
Internal data analysis and insights
Consequence to people: MediumAnalysing business data, surfacing insights, building dashboards
Workflow automation across systems
Consequence to people: MediumAutomating tasks across business systems - triggers, data movement, decisions in workflows
Governing Claude.ai in practice
- Name one person who owns Claude.ai in your business. Governance starts with a name.
- Write down what Claude.ai is used for. Certrak recognises 7 common uses: General productivity assistance, Document analysis and review, Marketing content creation, Customer communications drafting, Code generation and review, Internal data analysis and insights and Workflow automation across systems. Yours may differ; write down yours.
- Decide what Claude.ai may and may not be given. Personal information is in scope for some uses; write the boundary down before it is needed.
- There is no human review of outputs. Decide whether that is acceptable for each use, and what would make you change it.
- Claude.ai can create, change, or delete records in another system without a person approving. Record which of these you allow, and who can change that.
- Look at this again every month, and sooner on any of these: industry incidents, vendor model updates, autonomy level changes and data classification changes.
What changed, and how Certrak responded
We record every check here, including the ones that changed nothing. A check that found no governance change proves we looked.
-
Governance change Vendor change
Claude Cowork and chat are one Claude
Anthropic shipped it 16 Sep 2026.
Reviewed the app release notes from 3 September to 7 October. Anthropic has merged Cowork into every conversation: what was only possible in Cowork, including working in local files, browsing, computer use through the desktop app and scheduled tasks, is now available from any chat, with an Auto or Manual permission setting per conversation. Pro and Max first; Enterprise administrators get 30 days notice. Claude Cowork no longer has its own page here.
How Certrak responded: Where Claude is used for everyday work, documents or internal analysis, it can now change files and act in connected tools from an ordinary conversation. Owners are asked whether their people use Claude this way; if they do, the system is treated as making changes on their behalf and the matching controls apply.
-
No governance change Scheduled review
Claude release notes
Anthropic shipped it 28 Aug 2026.
Reviewed the recent release notes against Claude.ai's use patterns. No change to what it can do without approval, the data it is expected to touch, or who reviews its output.