Skip to content
Limestone AI Labs

AI for IT operations

Governed AI tools that work across your whole stack. Claude connected to your systems through MCP, read-only by default, every call logged.

Scope

What the work covers.

  • MCP servers

    Your tenant, RMM, network, and backups available to Claude as tools.

  • Read-only by default

    Write actions added one at a time, each approved.

  • SSO & audit log

    Access through your identity provider, with every call recorded.

  • AI rollouts

    Claude, Claude Code, or ChatGPT set up for the team with clear rules.

Every project runs Discovery, Design, Build, Handover. How we work

Platforms

What we build on.

  • Anthropic Claude
  • Claude Code
  • Model Context Protocol
  • OpenAI

Hardware and licensing through MQ3 Labs, or bring your own.

Track record

Where we have done it.

  1. AI for IT operations
    Hundreds of governed AI tools for an enterprise IT team.

    Two dozen MCP servers, read-only by default, every call logged behind SSO.

Client names stay private. All selected work

Before you write

A useful first email covers four things.

  1. 01 The systems your team works in every day.
  2. 02 Your identity provider.
  3. 03 The questions that take your team longest to answer.
  4. 04 Your rules on what AI may and may not change.

Missing some of it? Send what you have; we will work out the rest together.

Questions

Scope questions.

Can the AI change things on its own?
Only what you approve. Tools start read-only; write actions are added deliberately and every call is logged.
Which AI do you use?
Claude from Anthropic and OpenAI's models, depending on what your team uses and allows.

Pricing, hardware, and other common questions

Get in touch

Start with a conversation.

Tell us about your project and timeline. Expect a written reply within one business day, from the engineer who will do the work.

Prefer email?
hello@limestoneailabs.com
Existing clients
Sign in to the client portal →