Skip to content
The State of Business Security in TexasGet it free

Just published · White paper

The State of Business Security in Texas

A clear, honest guide for the leaders who don't have time for jargon: where Texas business security stands today, the five ways businesses actually get hurt, and what to do in the next 90 days.

Free PDF · AstraLink Connect

Get the white paper

See What Your Team Is Really Doing With AI

AI Protect gives you a plain-English view of how AI tools are being used across your business, and catches client data at the moment it's pasted. Watch it happen below.

Get a free AI risk checkTalk to our team

Animated example. A team member copies details from a client's tax return and pastes them into an AI chat. AI Protect checks the paste on their computer, finds a Social Security number and a bank account number, and puts a redacted copy on the clipboard. They keep the redacted copy, or paste the original and that choice is recorded for review.

Example with made-up names and numbers.

More than 8 in 10 employees use AI tools nobody signed off on.

Source: Teramind, Shadow AI Behavior Report

That's not a guess. It's what research on workplace AI use found, and it holds up across just about every business we've looked at.

If that sounds like your business, you're in good company. Most owners have no idea what's actually happening. That's exactly what AI Protect is for.

See every AI tool in use.Including the ones nobody signed off on.

A straightforward list of which AI tools your team is actually using. No more guessing what's happening on the other side of the screen.

One list, in plain English

Who's using what, and whether you've approved it.

AI tools in useLast 7 days
ToolPeopleUsed from
ChatGPT, personal accounts9Browser
Microsoft Copilot6Desktop app
AI meeting notetaker4Desktop app
Writing assistant extension3Browser

Example view with sample data.

No asking around the office

Most owners don't know what their team is using. That's normal, and it's the first thing AI Protect answers.

This weekSummary
AI tools seen4
Approved by you1
Waiting for your call3

Example view with sample data.

Stop it at the paste.Before it goes anywhere.

When someone is about to paste something they shouldn't, a client's tax return, a patient's chart, a case file, a customer's account number, AI Protect catches it first.

You decide what counts as sensitive

Client records, financial details, medical information. It looks for what you tell it to.

Watching forYour office
  • OnSocial Security numbers
  • OnBank and card numbers
  • OnPatient and medical details
  • OffClient names

Example settings.

The work still gets done

The risky part is held back. Everything else goes through, so nobody's stuck.

Recent catchesToday
  • Tax return fieldsCoached
  • Bank account numberBlocked
  • Patient date of birthRedacted

Example view with sample data.

Steer, don't shut down.Keep working fast, without the risk.

Instead of banning AI, AI Protect points your team toward doing it the right way, so they keep moving without putting the business on the hook.

A nudge, not a lecture

People see why something was held and what to do instead.

AI Protect

This looks like client work. Your office has an approved AI workspace for client documents.

Example message.

No slowdown

It runs quietly in the background. Your team keeps the AI tools that help them get things done.

Your team today
  • Drafted engagement letterAllowed
  • Summary email, details removedRedacted
  • Newsletter ideasAllowed

Example view with sample data.

See AI Protect in action

Same product, four very different Tuesdays.

Pick the one that sounds like your office.

What happens

What AI Protect does

Talk to someone who knows your industry

Does it read everything my staff types?

No. It looks for the data you named, at the moment of a paste or upload into an AI tool, on the computer itself.

Nothing is sent to AstraLink, or anyone, to be scanned.

  1. Never leaves the computer

    • The text or file being pasted or uploaded
    • Keystrokes, screens, browsing history
    • Anything typed that is not a paste into an AI tool

  2. What the console receives

    • Which kind of sensitive data was found
    • Who, which AI tool, and what AI Protect did
    • A masked snippet of surrounding textOnly if you turn it on. Off by default.
    • Which AI assistant and model a person usedOnly for people you opt in, one by one.
Read the AI Protect privacy policy

Why owners choose it.Built for the person actually running the place.

You do not have an IT department.

AI Protect installs from a link you email. Your MSP can run it for you from the same console.

You do not want to ban AI outright.

Your team keeps the tools. The client data stays home. Nobody gets pushed to their personal phone.

You want a straight answer.

One feed in plain language: who, what kind of data, which tool, what we did. Not a dashboard full of acronyms.

Whichever AI they open.AI Protect is already there.

Chat assistants, copilots, notetakers and writing extensions. Client data is caught at the paste, before it reaches any of them.

  • ChatGPT
  • Claude
  • Notion
  • Gemini
  • Perplexity
  • Midjourney
  • Cursor
  • DeepSeek
  • Grammarly
  • Grok
  • Kimi
  • ElevenLabs
  • Bolt
  • Lovable
  • Gamma
  • Replit
  • Devin
  • Luma
  • GitLab
  • Claude Console
  • Ollama
  • Fireflies
  • Jasper
  • Krea
  • QuillBot
  • GPT4All
  • AnythingLLM

For your IT person or MSP

The details they will ask about.

Short answers. The full admin reference is in the documentation.

Platforms. Apple Silicon Macs on macOS 13 and later. 64-bit Windows 10 (19041 and later) and Windows 11. Same rules, same console, both platforms.

Browsers. Chrome, Edge and Brave through a Chrome Web Store extension, which you can force-install by policy. Uploads and web traffic are also read by the inspection proxy on the device.

Install and updates. One emailed link per employee, or the same signed, notarised installer pushed by Jamf, Intune, Kandji or Hexnode. MDM is never required. The agent updates itself in staged cohorts, and updates can be pinned or disabled per tenant.

Policy model. Data type × destination × channel → allow, log only, coach, redact or block. First matching rule wins. Policy is assigned to the person and follows them across their devices. Changes reach devices within one check-in.

Detection. 22 built-in validated data types (SSN, ITIN, passport, driver's license, card and bank numbers, routing numbers, IBAN, EIN, crypto wallets, API keys, health IDs such as NPI and Medicare number), plus your own lists and patterns. US identifiers. Two engines, both on the device: pattern matching with validators, and a name and context recogniser.

Console. Multi-tenant with strict isolation, so an MSP runs several firms from one login. Per-device status: proxy, certificate trust tier, classifier health, agent version. Quiet devices show Unknown, never a green they have not earned.

Reflects the console and agent as deployed in October 2026. Documentation and admin reference.

Straight answers.

Something else on your mind? Ask our team.

Does AI Protect stop my team from using AI altogether?

No. This isn't about shutting AI down. It's about knowing what's happening and catching the risky stuff before it leaves your business. Your team keeps working the way they already do.

How does it know what is actually risky?

AI Protect watches for the kind of information you'd never want to leave your business by accident: client records, financial details, medical information, anything sensitive. When it sees something like that headed for an AI tool, it steps in.

Do I have to block everything right away?

Your call. AI Protect can run quietly in the background so you can see what's happening first, or it can step in and stop things in real time. Most owners start by watching, then tighten things up.

Will this slow my team down?

No. It works quietly in the background. Your team keeps using the AI tools that help them get things done. They just won't be able to accidentally hand over something they shouldn't.

What if I don't know what AI tools my team is using?

That's normal. Most owners don't. That's exactly why AI Protect exists, to give you that picture without you having to ask around the office.

Does this read everything my staff types?

No, and it's not a productivity monitor. It looks for the specific kinds of sensitive information you tell it to look for, at the moment something is pasted into an AI tool. Checking happens on the machine itself.

Ready to see what's actually happening?No pressure. No jargon.

Just a plain conversation about what AI Protect would catch in your business.

Get a free AI risk checkTalk to our team