Questions and answers.

It shows which AI applications and websites are used across your enrolled devices, lets you allow or block them by company, site, department, or device, and keeps reports and an audit history of device status and policy changes for your reviews.

The agent runs on 64-bit Windows 10 and Windows 11. macOS is not supported yet.

It checks the device against our list of known AI services: the names of programs that are running, the website addresses the device has recently looked up, and which AI apps from the list are installed. Only matches with the list are reported. Everything else stays on the device and is never sent to us.

Yes, from agent version 1.3.2. Every few hours the agent checks which AI apps from our list are installed, including per-user apps, Microsoft Store apps, and command-line tools, and the portal shows where each one is installed separately from where it is used. Only matches with our list are reported; nothing about other software leaves the device.

No tool can promise that. Detection covers the services on our list, which we keep updating. A browser set to use its own secure DNS can hide website visits from the device, and a program is only seen as in use while it is running (installed AI apps from our list are reported separately). It gives a strong practical view of AI use on enrolled devices, not a guarantee.

When a policy blocks a website, the agent stops that site loading on the device and shows a page in that browser tab explaining that your organisation's AI policy blocks it. When it blocks an AI program, the agent closes it within seconds of it starting and shows a notice on screen. Policies reach each device at its next check-in. Check-ins are every 15 minutes by default and company owners can choose an interval from 5 minutes to 2 hours. Removing the agent removes the blocking.

Yes, by default, so that a blocked AI website can show its explanation in the same browser tab. Each device creates its own certificate, trusted on that device only, with a private key that is created on the device and cannot be exported. The agent uses it only for AI websites your policy blocks: it answers those connections itself, so nothing is sent to the blocked website, and connections to other websites are not opened or read. Company owners can turn this off on the Security page in the customer portal, which removes the certificate at each device's next check-in; the explanation then opens in a new tab instead. Removing the agent also removes the certificate.

Every company starts with a report-only policy: AI tools keep working and are reported in the portal, and nothing is blocked. When you are ready, an owner approves or blocks tools and publishes your own policy, which replaces it.

Yes, by default. An AI Assure icon sits in the system tray. Its window shows whether the agent is working, your organisation's AI policy (which tools are allowed and which are blocked), anything blocked recently, and what is reported about the device and what is never recorded. When something is blocked, people see a short notice that names it, with a link to ask IT for access if your organisation sets a contact. Company owners can turn the icon off and set who to contact on the Security page.

The agent keeps applying the last policy it received, then checks in again when it reconnects and picks up any newer policy. Company owners can set when offline-agent alerts are raised; the default delay is 48 hours to account for devices that are switched off outside working hours.

Yes, from agent version 1.3.0. When we release a new version, each device downloads it at its next check-in, checks it, and installs it quietly, keeping its registration, policy, and settings. Devices on an earlier version need the current installer run once from Devices in the customer portal. What changed in each version is listed under Agent releases in the portal.

No. AI use is reported at service or website level: which service, on which device, and when. The agent does not read prompts, typed content, uploaded files, AI responses, page content, or the screen. See Trust and privacy.

No. There is no remote access. The agent only makes outgoing secure (HTTPS) connections to AI Assure: to register, check in, collect its policy, and report what it found. We cannot connect in to a device, run commands, see its screen, or open its files, and the agent does not accept connections from other computers: its local web filter listens only on the device itself. The only software it installs is its own updates, which we publish and it checks before installing.

The device name, a Windows machine identifier, the operating system and agent version, check-in times, the names of AI services found with the time they were seen, which AI apps from our list are installed (open or not), programs closed because of your policy, and a fingerprint of the installed programs that shows when they change without sending the list itself.

In the United Kingdom. AI Assure runs on our own servers in our data centre in the UK, and customer data is stored there. The providers we use, such as for payment setup, are listed on our Subprocessors page.

Yes. People in your company are owners, standard users, or read-only users, and an owner can limit a person to particular sites and departments so they only see those devices and reports. Two-step sign-in is required for every account, and the portal signs people out after a period of inactivity that owners choose.

Priced per active agent each month, plus VAT. There is no minimum quantity. Use the pricing calculator to work out your monthly total.

Each month you receive an invoice by email, in advance for the month ahead. Any account credit is taken off first, and the rest is collected using the payment method shown at sign-up. You set up payment on our payment provider's secure page during sign-up, and AI Assure does not store your card or bank details.

Changes to the number of agents apply from your next billing period, and the portal shows the new monthly total before you confirm. Nothing is prorated: the current period stays as it is.

We email the account owners straight away and try the payment again. If it is still unpaid after seven days, the account is limited to billing until it is paid, with a final notice the day before. While limited, agents keep applying their last policy but stop reporting. Full access returns automatically once the invoice is paid.

The account owner can request cancellation from Billing in the customer portal. We confirm the end date by email. The account ends at the end of the current paid billing period, no further invoices are raised, and the agent uninstalls itself from each device at its next check-in. See Billing and Cancellation.

After the owner's email address is verified and payment setup is complete. Until then, nothing is activated or charged.

In the customer portal, open Devices and choose Add devices. Pick the site and department, then download the Windows installer and enter the key it shows. You can install it by hand or silently through your device management tool. Each device registers itself and appears in the portal within a few minutes.

Go to the customer portal and use your company URL, email address, and password, then the code from your authenticator app, as two-step sign-in is required. If you have forgotten your company URL, choose Forgot your company URL? on the sign-in page and we will email it to you. If you have forgotten your password, reset it here.

Use the contact form. Existing customers can also send a request from Support in the customer portal.