Most co-managed IT arrangements fail for a simple reason: nobody wrote down who does what. The internal team assumes the provider is watching alerts. The provider assumes the internal team is approving patches. Both are wrong, and the gap shows up during an outage.
Co-managed IT means your internal staff stays and an outside provider shares the workload. Z7’s managed IT page describes it this way: your internal team stays, and Z7 adds 24/7 coverage, enterprise tooling, and specialist depth beside them. The model works well when the division of labor is explicit.
Decide the model before the details
Co-managed fits when you have capable people but not enough hours, coverage, or specialists. Typical pressures are after-hours alerts that fall on one person, a help desk backlog that crowds out projects, and security tooling nobody has time to run.
Fully managed fits better when there is no internal IT, when the one person who does it is leaving, or when the organization wants a single accountable party for the whole function. Z7 describes fully managed as owning the help desk end to end, along with monitoring, patching, security, vendors, and strategy. A co-managed arrangement with no internal owner is really a fully managed one under another name, and it should be scoped that way.
Build a responsibility matrix
Write one row per function and one column per party. A common format is RACI: who is Responsible, who is Accountable, who must be Consulted, and who is Informed. CISA’s guidance for customers of managed service providers recommends that contracts state whether the provider or the customer owns specific responsibilities, such as hardening, detection, and incident response, and that a shared responsibility matrix be documented.
Here is an illustrative starting point. Your own split will differ, and the signed agreement controls.
| Function | Internal team | Provider | Notes to settle |
|---|---|---|---|
| First-line help desk | Consulted | Responsible | Which hours, which channels |
| Escalated user issues | Responsible | Consulted | Handoff trigger and contact |
| Server and network monitoring | Informed | Responsible | Alert routing, after-hours calls |
| Patching | Accountable for approval | Responsible for execution | Approval windows, exceptions |
| Backup monitoring | Informed | Responsible | Who tests restores |
| Security detection and response | Accountable for business decisions | Responsible for investigation | Containment authority |
| User onboarding and offboarding | Accountable for HR trigger | Responsible for execution | Approval workflow |
| Vendor and license management | Responsible | Consulted | Who talks to each vendor |
| Projects and upgrades | Accountable | Responsible if scoped | Written scope per project |
| Strategy and budgeting | Accountable | Consulted | Review cadence |
The “notes” column matters most. Each note is a decision someone has to make before go-live.
Agree on tooling and access
Shared tools need shared rules. Settle these early:
- Who administers the remote monitoring and management platform, and can your staff see everything the provider sees?
- Does your team get its own accounts in the ticketing system, security tools, and backup console, with the access level its role requires?
- Who owns the data and configuration if the relationship ends?
- How is provider access to your environment secured? CISA recommends requiring MFA on all provider accounts used to access customer environments, not assigning provider accounts to internal administrator groups, and not reusing administrator credentials across customers.
- How is access reviewed and removed when people change roles or contracts end?
Insist that your administrators keep independent access. A provider-controlled tenant with no internal owner is a continuity risk.
Define escalation both ways
Escalation runs in two directions. The provider needs a named contact on your side for decisions, approvals, and after-hours emergencies. Your staff needs a documented route to the provider’s senior engineers and a management contact for stalled work.
Write down the triggers: severity level, time without progress, or a security signal. Z7’s help desk services and published NOC description reflect this idea: monitoring with alerting and clear escalation so the right issue reaches the right person. Whatever provider you pick, ask to see the path and test it once with a non-urgent issue.
Separate run work from project work
Run work keeps the lights on: tickets, monitoring, patching, backups. Project work changes the environment: migrations, a new firewall, a tenant consolidation. Co-managed models often fail by blurring the two.
A reasonable rule is that the provider takes the repeatable run work, which frees your staff for projects. Z7 describes its NOC role in co-managed engagements the same way: the provider takes the monitoring and the after-hours load while your staff works on projects. Confirm in writing which projects are included, which are scoped separately, and how change approvals work. Exclusions depend on the proposal, so check projects, hardware, licenses, after-hours work, onsite requirements, and third-party costs before signing.
Set a reporting cadence
Agree on what you will see and when. A workable rhythm:
- Weekly: a short operational check-in covering open tickets, upcoming changes, and anything urgent
- Monthly: a report on ticket volume, performance against service targets, patch status, backup results, and recurring issues
- Quarterly: a business review covering roadmap, risk, budget, and the responsibility matrix itself
Revisit the matrix at each quarterly review. Staff turnover, new systems, and new compliance duties all shift ownership, and the document should keep pace.
Warning signs that co-managed is not working
- Your staff cannot see what the provider sees
- Tickets bounce between teams without an owner
- After-hours incidents still wake the same person
- The matrix has not been updated in a year
- Project requests keep landing in a gray zone
If several apply, tighten the matrix or consider moving more of the function to the provider.
Where Z7 fits
Z7 offers managed IT as co-managed or fully managed, with the NOC and help desk as building blocks. If you want to draft a responsibility matrix for your environment, book a conversation.