AI governance for MSPs
See the AI your clients already use.
Install the browser extension and the Windows agent, and Veritus builds an AI inventory for each client: which tools, on how many computers, and which ones nobody has reviewed. Then set the rules and keep the evidence.
14 days free. Card required. Your plan starts billing after the trial unless you cancel in Billing before it ends.
Sample Dental Group
24 computers reporting
- AI tools found
- 9
- Computers using AI
- 17 of 24
- Not reviewed yet
- 6
| AI tool | Seen in | Computers | Review |
|---|---|---|---|
| ChatGPT | Browser | 14 | Not reviewed |
| Microsoft Copilot | Browser | 11 | Approved |
| Cursor | Desktop app | 3 | Not reviewed |
| DeepSeek | Browser | 2 | Not approved |
5 more AI tools
Evidence for Q3 4,812 prompts and uploads checked, 137 warned, 22 blocked
Check prompts and supported uploads before they are sent
Policies look for personal data, health information, payment data, secrets, source code and your own keywords. Pick a response for each policy and simulate it before you enforce it. People see why, in plain words.
chatgpt.comExample
Write a friendly payment reminder for Dana Whitfield.Card on file: 4111 1111 1111 1111Email: dana.whitfield@example.com
Sent as written. Nothing was shown to the person.
- Found
- Credit card, email address
- Policy
- Payment data v2, monitor
- Outcome
- Sent as written
chatgpt.comExample
Write a friendly payment reminder for Dana Whitfield.Card on file: 4111 1111 1111 1111Email: dana.whitfield@example.com
Sensitive data detected
Credit card, Email. Proceed?
- Found
- Credit card, email address
- Policy
- Payment data v2, warn
- Outcome
- Warned, then edited
chatgpt.comExample
Write a friendly payment reminder for Dana Whitfield.Card on file: 4111 1111 1111 1111 [redacted-credit-card]Email: dana.whitfield@example.com [redacted-email]
Removed 2 sensitive values (Credit card, Email). Your message was sent without them.
- Found
- Credit card, email address
- Policy
- Payment data v2, redact
- Outcome
- Removed, then sent
chatgpt.comExample
Write a friendly payment reminder for Dana Whitfield.Card on file: 4111 1111 1111 1111Email: dana.whitfield@example.com
Justification required
This prompt requires a justification before it can proceed (Credit card, Email).
Practice manager approved this reminder.
- Found
- Credit card, email address
- Policy
- Payment data v2, ask for a reason
- Outcome
- Sent with a reason
chatgpt.comExample
Write a friendly payment reminder for Dana Whitfield.Card on file: 4111 1111 1111 1111Email: dana.whitfield@example.com
Blocked by your organization policy: Credit card, Email.
- Found
- Credit card, email address
- Policy
- Payment data v2, block
- Outcome
- Blocked, not sent
Evidence a client can hand to an auditor
Veritus records what happened, not just what the policy said: the AI site, what was detected, the decision, what the person did next and the policy version in force. Veritus also shows which browsers and computers stopped reporting, so nobody mistakes “not observed” for “didn't happen”.
- Mapped to ISO/IEC 42001, ISO/IEC 27001, NIST SP 800-171 and CMMC practice IDs
- Any client, any date range, as PDF or CSV
- Reviewer notes on decisions that should not have triggered
Veritus maps AI activity to framework controls to support your audits. It does not certify compliance.
- Event
- EVT-20481
- Client
- Sample Dental Group
- AI site
- chatgpt.com
- Found
- Patient name, date of birth
- Policy
- PHI v4, warn
- Outcome
- Warned, then cancelled by the user
- Controls
- ISO 27001 A.5.34, NIST 800-171 3.3.1, CMMC AU.L2-3.3.1
- Recorded
- 18 Sep 2026, 14:32 EDT
Integrity hash sha256 9f3c 41d2 07be a71e
| Prompts and uploads checked | 4,812 |
|---|---|
| Warned | 137 |
| Blocked | 22 |
| Sent with a reason | 9 |
| AI sites checked in the browser | 5 |
| Framework requirements with evidence | 14 |
Counts are policy decisions in the browser, not unique people. Every report states its sources and what it leaves out.
Built for the way MSPs work
One console for every client
Switch between clients, and see on one home screen which ones have computers or browsers that stopped reporting.
Policy bundles across clients
Build a policy set once, with the frameworks it maps to, and apply it to any client in a few clicks.
Deploy with the tools you have
Send a Windows installation invitation by email, or install silently through Intune or your RMM. The agent also installs the browser extension, locked so staff cannot remove it.
A service you can sell
Discovery findings and a monthly evidence report give you a recurring deliverable, not a one-off project.
Your clients
12 clients 11 reporting 1 needs attention
Sample Dental Group
Reporting
4,812 prompts checked
Harbor Street Law
Reporting
2,140 prompts checked
Northside Family Clinic
Reporting
3,377 prompts checked
Keel and Ward Accounting
Reporting
1,906 prompts checked
Brightline Logistics
Reporting
988 prompts checked
Cedar Point Realty
Reporting
642 prompts checked
Oakhurst Veterinary
Reporting
1,215 prompts checked
Summit Civil Engineering
Reporting
2,731 prompts checked
Riverbend Credit Union
Reporting
5,094 prompts checked
Lakeview Insurance Agency
Reporting
1,488 prompts checked
Pinecrest Orthodontics
Needs attention
2 computers stopped reporting
Maple Ridge Physical Therapy
Reporting
1,102 prompts checked
Open the conversation with a free AI risk assessment
Most businesses don't know which AI tools their staff use. Show a prospect in two weeks, then propose the service that fixes it.
Run your first assessment free- Day 1
Add the prospect as an organization and send the install link. Monitor mode: nothing is blocked and nobody is interrupted.
- Days 2 to 14
Veritus finds the AI tools staff use, on which computers, and the sensitive data that shows up in prompts and uploads.
- Day 15
Hand over the client report: what was found, what it means, and which framework controls it touches. Keep the organization if they sign.
Bans don't stop AI use. Policies on paper don't prove anything.
Your clients' staff already paste client data into AI tools. When their auditor, insurer or customer asks how that is controlled, the usual answers fall short.
Block every AI site
Staff switch to personal phones and accounts. You lose sight of the use and still have nothing to show an auditor.
Send an acceptable-use policy
Everyone signs it. Nobody can show it was followed, or what was pasted into an AI tool last month.
Buy an enterprise AI security tool
Built for one company's security team, often with seat minimums of 50 to 1,000 and no view across your clients.
Choose how you run AI governance
Run it yourself, get help launching your MSP, or arrange ongoing management with Cyber Runtime.
Self-service
Operate AI governance with your own team.
$99.00 per month per MSP
25 users included.
$2.50 per additional user/month.
Card required · $0 due today
- AI inventory across your clients
- Browser prompt and supported-upload policies
- Policy simulation and reusable bundles
- Framework-mapped PDF and CSV evidence reports
- Windows discovery and email installation invitations
Guided MSP launch
Launch your first client rollouts with practical help.
$499.00 per month per MSP
100 users included.
$2.50 per additional user/month.
Card required · $0 due today
- All Self-service software capabilities
- Two rollout sessions
- Policy configuration for your first two clients
- One evidence review each quarter
- Software trial first; guided services start after payment
Managed governance
Planned ongoing governance with Cyber Runtime.
$1,500.00 per month per client
100 users included.
$2.50 per additional user/month.
Coming soon · No card or payment collected
- All current Veritus software capabilities
- Management for one client organization
- Human review and policy maintenance
- Findings follow-up and management reporting
- Delivery agreement being finalized
Prices are monthly, in US dollars, before tax. Self-service and Guided MSP launch include a 14-day software trial with a card required. Guided services begin after the first payment. Cancel in Billing before your trial ends to avoid a charge. Existing enterprise pilot agreements remain unchanged. See the service scope.
Coverage, stated plainly
Better you know the limits before a pilot than find them during one.
| Where | What Veritus does | Status |
|---|---|---|
| Chrome and Edge on ChatGPT, Claude, Gemini, Copilot (web), Perplexity and DeepSeek | Checks prompts and supported file uploads. Monitor, warn, ask for a reason or block. | Available |
| Chrome and Edge on about 100 other AI tools, such as Grok, Midjourney and Otter.ai | Blocks the website of a tool the client has not approved, with a page where the person can ask for access. Prompts are not inspected. | Available |
| Windows 10 and 11 computers | Finds installed AI apps, AI programs that run, visits to about 500 AI websites, MCP servers and AI browser extensions. Optional controls close AI apps the client has not approved, turn off private browsing and turn off the AI built into Chrome and Edge. | Available |
| Desktop AI apps, command-line tools and IDEs | Discovered by the Windows agent. Prompts are not inspected. With agent controls on, an AI app the client has not approved is closed, and the person can ask for access. | Discovery and controls |
| macOS, Firefox, Safari, Microsoft 365 Copilot | Not covered today. | Not yet |
Browser checks require an enrolled, connected extension. Upload coverage depends on file type, size and readable text; unsupported files and unreadable content are not fully inspected. The Windows agent records app launches and network connections from every program, including system tools such as PowerShell, so it can tell which ones use AI. It never records what is typed or shown. Nothing is blocked or closed until you mark a tool Not approved and choose to enforce, and agent controls start with a few test computers. See what Veritus collects, where it is stored and how long it is kept.
Start with one client. Try Veritus free for 14 days.
Prove the value on one client before you roll it out to the rest.
Start your 14-day free trial- Day 1
Choose a plan and add a card to start your 14-day free trial. Install the browser extension and Windows agent for one client.
- Days 2 to 14
See which AI tools are in use, review the evidence and try your policies in monitor mode before enforcing them.
- After 14 days
Your selected plan starts billing automatically. Cancel in Billing before the trial ends to avoid a charge.