AI governance for modern teams

Stop approving AI tools in scattered Slack threads and emails.

One place to request, review, and approve every AI tool your team uses. GovernKit checks each vendor against its own terms and keeps the record for you.

Get started with the free plan. No credit card required.

Northstar StudioOwner
DashboardAI ToolsRequests1PoliciesReportsMembersAudit LogSettings
Business planManage
12 / 200 members
SCMaya Raomaya@northstar.co
AI Tools
Every AI tool your company tracks — approved, pending, and everything in between.
+ Add AI tool
ToolCategoryDepartmentsRiskStatus
CClaudeGeneral AI assistantEngineeringHighRestricted
MMicrosoft CopilotGeneral AI assistantMarketingLowApproved
GGrammarlyWriting assistantAll departmentsLowSubmitted
FFreeTranscribe.ioTranscriptionAll departmentsHighRejected
PPerplexitySearch/researchAll departmentsLowDeprecated
5 tools match · est. $358/month in tracked AI spend across approved tools
Why GovernKit

Your team should use AI. On your terms.

Your people already use AI, and banning it only hides the usage. The real risk is adoption without a record: unread terms, unclear data flows, approvals buried in chat.

Say yes faster

A clear request path beats a blanket ban or a forgotten DM.

Know what you're agreeing to

Every vendor is checked against its own terms before approval.

Grow without losing track

The shared directory, usage policy, and audit trail scale with you.

Approve AI tools with confidence

One flow takes a tool from request to documented decision.

1 · Tools in use
ChatGPT
Claude
Copilot
Cursor
Gemini
More tools
2 · Review
GovernKit
GovernKit
request · vet · decide
3 · Decision
Approved!With restrictionsRejected?More information
4 · Record
Company AI Policy
v6 · published
88% acknowledged
Audit trail
Request submittedReview in progressApprovedPolicy updated

A request goes in one end. A documented decision comes out the other.

How it works

Employees ask in two minutes. Admins decide with full context.

A structured request form replaces the ad-hoc DM.

1
A short form starts it

An employee picks a tool, explains the use case, and selects the data involved. The form also asks whether the tool is already in use, so your real inventory surfaces itself.

2
Vetting happens before review

GovernKit reads the vendor's documents and prepares a cited risk assessment before a reviewer opens the request.

3
The decision has four outcomes

Approve, approve with restrictions, reject, or ask for more information. A threaded Q&A lives on each request.

4
The record writes itself

The requester hears back by email, the shared directory updates, and the audit log gets another entry.

Request an AI tool
Takes about 2 minutes. You'll get an email as soon as there's a decision.
Which tool would you like to use?
ChatGPTClaudeGeminiCopilotSomething else…
What data will touch it?
Public informationInternal documentsSource codeCustomer data
Suggested: High riskupdates live
High risk because this request includes source code.
Your decision
ApproveEveryone can use it as requested.
Approve with restrictionsAllowed — but with rules everyone must follow.
RejectNot allowed. You'll explain why below.
Request more informationAsk the requester a question before deciding.
Risk level
High
Pre-filled with the suggested level — adjust if you know better.
Approve tool
The requester gets an email with your decision.
Automatic vendor vetting

We read the privacy policy so you don't have to.

Behind every review, GovernKit turns the vendor's privacy policy and terms of service into a color-coded risk table. Each flag carries a verified quote, so reviewers can see exactly where a concern came from.

Severity colors reviewers can scan at a glanceVerbatim citation behind every flagUnverifiable claims are labeled as “unknown”
Vendor policies & insights✦ AI-generated · awaiting review
Privacy policy Terms of service Website
Documents last checked Jul 7, 2026  ·  AI summary generated Jul 7, 2026
Risk assessmentMedium risk
CategorySeverityWhat the documents say
Data trainingMediumThe vendor says it may use user content to train its AI models, and users can control that training in account settings.
Intellectual propertyLowUsers own their outputs, and the vendor does not retain ownership of AI-generated outputs.
Data sharingMediumThe vendor discloses user content to trusted service providers for AI technology and cloud storage.
ComplianceLowThe privacy policy explicitly says the vendor uses encryption and other safeguards to secure data.
CautionData training

Flagged because the vendor's own policy says:

“User inputs may be used to train models unless users opt out through their account settings.”

Quote verified against the source text
Northstar AI Usage PolicyPublished · v6
Published Jul 6, 2026 by Maya Rao
Export PDF.md
Acknowledgements14 of 16 members (88%)
2 members haven't acknowledged yet.Send reminder to 2 members
Version history
Northstar AI Usage Policy · v5Published Jul 5, 2026ArchivedView
Northstar AI Usage Policy · v4Published Jul 5, 2026ArchivedView
Northstar AI Usage Policy · v3Published Jul 5, 2026ArchivedView
Policy generator

A real AI policy, plus proof your team read it.

Answer a short questionnaire and GovernKit drafts a workplace AI policy you can edit and publish. Acknowledgements are tracked, and reminders chase the stragglers for you.

AI-drafted, human-edited, versioned policiesAcknowledgement tracking with a live coverage percentageContractors and staff can sign by email invite, no account neededPDF and Markdown exports for clients, auditors, or insurers
Always current

Vendors change their terms. You will know when it matters.

GovernKit re-checks each vendor's documents on a schedule. When something changes, the risk assessment refreshes with new citations, your team gets an email, and the tool page shows when the vendor was last checked.

Vendor monitoringAutomatic
Documents last checked Jul 7, 2026 · next check scheduled
Vendor changed its privacy policyJul 7, 6:02 AM
Scheduled re-crawl detected a change in the vendor's privacy policy.
Risk assessment regeneratedJul 7, 6:04 AM
Data training flag raised from Low to Medium, with a new verbatim citation.
Team notified by emailJul 7, 6:05 AM
“A vendor behind one of your approved tools changed its privacy policy.”
Tools by status
20 tools tracked in total.
Approved8
Approved with restrictions3
Pending review2
Rejected4
Deprecated3
Monthly AI spend
$358estimated per month across approved tools
Microsoft Copilot$218/mo
Claude$96/mo
Grammarly$44/mo
1 request waiting more than a week — flagged as stale
Dashboard & reports

One page that shows whether AI is under control.

Reports break down tools by status and risk level and show where the money goes.

Monthly AI spend, ranked by toolStale requests and overdue queues are flagged automatically

Every decision, on the record.

When a client or insurer asks how you manage AI tools, you can answer with a link.

Explore all features
Two-factor authenticationQR setup and backup codesStripe-powered billingSaved cards and instant invoicesHonest plan limitsPlans match the limits inside the productNo prompt monitoringWe track approvals, never your people

Questions, answered plainly.

Minutes. Create a team, invite your people by email, and share the request form. Everything is guided, and there is nothing to install.

Know every AI tool your team uses.

Share the request form with your team and see the first documented decision today.

Free plan · cancel anytime