Lansafe AI Studio
News

LanGuard goes live: our first AI Studio product, built on Lansafe hardware and now running at NWMS

· 3 min read · Dave Hayward

A laptop glowing with a pink shield in a factory office at dawn, a metal workshop beyond the window

LanGuard by Lansafe is now live. It is the first Lansafe AI Studio product to move from idea to real, everyday work: an always-on security watch for Microsoft 365, the firewall and remote access that explains what happened overnight in a short, plain-English email every morning.

It was built and proven on Lansafe's own hardware, and it now runs on equipment owned by NW Metal Sections (NWMS), a UK manufacturer. Same software, new home. That is exactly what we mean when we say everything we build is free to come home.

What LanGuard does

  • Identity watch. Every hour it reviews Microsoft 365 sign-ins, failed logins, mail sent from unfamiliar connections, inbox rule and forwarding changes, and app consents.

  • Network watch. Gateway and internet line status, intrusion prevention mode, blocked attacks, devices online, firmware and admin logins.

  • Remote access watch. Who connected, for how long, and any failed or unusual logins.

  • A morning report people actually read. All clear, or items to review, each with a recommended next step. Urgent problems raise an alert straight away.

The first week, in numbers

Across five daily reports in early November, LanGuard counted and explained:

  • 71 intrusion attempts stopped at the firewall: every one blocked, none merely detected.

  • 457 successful Microsoft 365 sign-ins reviewed, all from the office or a known home connection.

  • 41 failed sign-ins triaged. All came from inside the office and were dismissed, with the reason written down, as mistyped passwords.

  • 2 items escalated to a person, each with a clear fix.

  • 0 failed remote-access or firewall admin logins.

One escalation is worth telling. Overnight, the firewall vendor's service quietly stopped letting LanGuard read the security logs. Most tools would simply go quiet. LanGuard raised an alert straight away, confirmed it every hour, and the morning report gave the exact fix. A watchman who tells you when he can't see is worth having.

We have deliberately left out addresses, account names, device names and system details. A security report is a map of a network, and maps don't belong on a news page.

Rules first, AI where it earns its place

LanGuard is a good example of how Lansafe AI Studio works. Scripted, tested checks gather and count the data the same way every hour. AI is used for the part that needs judgement: deciding what is unusual and explaining why. Every finding is marked dismissed or escalated with a reason, so the IT team can check the thinking rather than read raw logs. No per-seat licences, no new dashboards, no agents installed on staff PCs.

Built here, running there

LanGuard was developed and tested on Lansafe hardware. When it was ready, it moved onto NWMS's own equipment, with its rules and records intact. It does not have to live there: the same service can run on Lansafe's hosting, in a customer's cloud, or on a server in their building. The customer decides where their security data lives, which for many businesses is becoming a compliance question, not just a preference.

You can meet LanGuard on our Digital Workforce page, read the full portability case study, or see the LanGuard service page.

Could your business do the same?

If one person looks after all your IT and nobody has time to read logs, LanGuard was built for you. And if you have a different process that eats hours every week, tell us about it or get in touch.