Claude, ChatGPT and Copilot, set up properly and connected to your Microsoft 365
Most organizations end up with more than one AI tool. We help you decide which tool does what, roll it out safely, connect it to SharePoint, Outlook and Teams, and build agents that remove real work. We don’t resell licences, so our advice is about fit, not commission.
- Your dataSharePoint · Outlook · Teams · OneDrive
- What we set upMindlab governance layerWho can use what · Same permissions as today · Monthly spend caps · Evidence for your risk team
- Any AIMicrosoft Copilot · Claude · ChatGPT
What we do on each platform
Claude by Anthropic
- Roll out Claude Team or Enterprise with single sign-on, admin controls and usage policies
- Connect Claude to SharePoint, OneDrive, Outlook and Teams with Anthropic’s Microsoft 365 connector, using the permissions people already have
- Build custom connectors (MCP) to your own systems, and Claude-powered apps on the API or in Microsoft Foundry
- Set up Claude Code for your developers, with review and security guardrails
ChatGPT by OpenAI
- Roll out ChatGPT Business or Enterprise with single sign-on, workspace settings and data controls
- Connect company knowledge from SharePoint, Outlook and Teams, with the Microsoft Entra permissions approved properly
- Design custom GPTs for repeat tasks, with owners and review dates
- Build apps and agents on OpenAI models, directly or through Azure OpenAI in Microsoft Foundry
Microsoft Copilot
- Plan and roll out Microsoft 365 Copilot, including fixing SharePoint oversharing first
- Build Copilot Studio agents and Power Platform workflows, from pilot to production
- Choose the right model per agent: OpenAI or Anthropic Claude models are both available in Copilot Studio
- Control cost with per-agent credit limits and budgets
What every rollout includes, whichever tool you pick
Before you switch it on
- Which tool for which job, written down in a one-page AI use policy
- Single sign-on with Microsoft Entra ID, and who can use what
- SharePoint and Teams permissions checked, so AI only finds what people should see
- Data settings reviewed: retention, training opt-out, where data is stored
After go-live
- Hands-on training built around your own work, not generic demos
- A short list of approved prompts, GPTs, projects and agents, each with an owner
- Usage and cost reviewed every month against the value it creates
- Updates when Microsoft, Anthropic or OpenAI change something that affects you
Who we work with
I take on a limited number of clients at a time, and work directly with each one.
Mid-size organizations (our focus)
A plan for which AI tool does what, connected to your files and systems, with one or two agents that remove real manual work.
Regulated and audited work
Governance your risk team will sign off: identity, data access, release process, evidence and cost control across every AI tool.
Small teams
One tool set up safely, a one-page AI policy, and your team trained in half a day. About two weeks.
Not sure which AI tool fits?
Tell us what your teams spend their time on. In 30 minutes you’ll know which tool fits each job, what it would cost, and what to set up first.