An AI management checklist for managed service providers
Clients increasingly expect their managed service provider to help with AI, not just email and devices. The good news is that the work is familiar: know what is in use, control access, watch costs and review regularly. This checklist turns that into a repeatable service for each client.
Onboarding a client
- Build a register of the AI tools the client uses, with purpose, teams and who looks after each one.
- Find which apps people have connected to the client's Google Workspace or Microsoft 365 accounts.
- Agree which tools are approved and how staff ask for new ones.
- Note renewal dates and how each tool bills.
Access and security
- Check company sign-in and two-step verification for the accounts used with AI tools.
- List API keys, their scope and last use. Retire keys that are unused or open to the whole organisation without a reason.
- Make sure AI tools are part of the client's leaver process.
Spend
- Review seat counts against real use each month and before renewals.
- Set spend limits and alerts on API accounts.
- Report seat subscriptions and API usage as separate lines.
Running it as a service
- Keep each client's records completely separate.
- Review each client quarterly, and after any leaver or renewal.
- Give each client a short summary: what changed, what needs a decision and what you did.
Common questions
Should an MSP manage a client's AI tools?
Many clients already expect it, and the work fits existing services: access, security, licences and reviews. Agree the scope with each client first.
How often should an MSP review a client's AI?
Quarterly works for most clients, plus a check whenever someone leaves or a renewal is due.