← Back to blog

MSP Onboarding Checklist: 8 Phases From Signed Contract to Go-Live

November 2, 2026
MSP Onboarding Checklist: 8 Phases From Signed Contract to Go-Live

A good MSP onboarding checklist moves a new client from signed agreement to stable, monitored service in eight phases: scope handoff, information and access collection, discovery, security baseline, service setup, training and communication, go-live, and a 30-day review. The order matters more than the tool you track it in. Access and inventory come before anything is installed, and nothing counts as done until it has been tested.

This guide is written for two readers: the business owner or operations lead who is bringing on a managed service provider, and the MSP that wants a repeatable process. Both benefit from the same list, because every item has an owner on one side and something to verify on the other.

Mindpodtech
Build a Stronger IT Foundation
Mindpod Technologies helps smaller companies plan practical technology, from security and resilience to cloud architecture and custom software.
Explore technology support

TL;DR

  • Onboarding is an operational handoff, not a software install. Plan it as a short project with named owners and dates.
  • Collect the client record and administrative access first, and store credentials in a vault, never in email or a spreadsheet.
  • Build the inventory during discovery, then check it against what your monitoring actually covers.
  • Go live only after you have tested alerts, ticket routing, remote access and a real backup restore.
  • Hold a 30-day review, remove the outgoing provider's access, and feed lessons back into the template.

Table of Contents

What MSP onboarding actually covers

Onboarding is everything between "yes, we'll work together" and "this runs as a normal service." It has a technical side (inventory, access, tooling, security baseline) and a human side (kickoff, communication, training, escalation paths). Most delays come from the human side: nobody knows who approves a change, who the after-hours contact is, or who holds the domain registrar login.

What MSP onboarding actually covers

The pages that rank for this topic all describe roughly the same arc, and so do we. What differs is how much weight each phase gets. We put discovery and security baseline earlier and heavier than most generic lists, because a provider that monitors an environment it never inspected inherits every hidden risk in it.

The eight phases at a glance

PhaseMain outputTypical owner
1. Sales handoff and scopeAgreed services, tier, exclusions, SLAAccount manager
2. Client record and accessContacts, sites, admin access in a vaultProject lead
3. DiscoveryInventory, network and cloud map, baseline risksTechnical lead
4. Security baselineMFA, least privilege, patching, backup testedTechnical lead
5. Service setupMonitoring, ticketing, documentation, billingService delivery
6. Communication and trainingKickoff, welcome kit, user trainingAccount manager
7. Go-liveCutover, hypercare, first-week watchService desk
8. 30-day reviewReconciled counts, remediation ownersAccount manager and client

Adjust the owners to your team size. In a five-person MSP one person may hold three of those roles. The point is that every row has exactly one name next to it.

Phase 1: Sales handoff and scope

The first failure point is the gap between what sales promised and what the technicians believe they must deliver. Before any work starts, put the following in writing:

  • The services in scope and the service tier, plus anything explicitly out of scope (line-of-business application support is a common surprise).
  • Response and resolution targets in the SLA, and the hours they apply to.
  • Who at the client can approve changes, spend, and emergency actions.
  • Known problems the client already told you about, such as an aging server or a failed audit.

If you are the business owner, ask to see this summary before the kickoff. If a provider cannot hand you one page that says what is covered, that is an early signal about how documentation will go later.

Phase 2: Client record and access collection

Start with a plain client record: legal company name and address, additional sites, number of users, main business contact, technical contact, hours of operation and emergency procedures. Then list the services the MSP will need to touch:

  • Domains, registrars, DNS and certificates
  • Internet providers and telephony
  • Microsoft 365 or Google Workspace tenants and other cloud services
  • Firewalls, switches and Wi-Fi
  • Line-of-business applications and their vendors, with support and renewal details

Access is where onboarding quietly goes wrong. Guidance for MSPs and their customers from CISA recommends restricting MSP accounts to the systems they actually manage, disabling provider accounts that are no longer used, and monitoring provider activity. Turn that into checklist items:

  1. Create named administrator accounts for each engineer instead of one shared login.
  2. Protect every administrator account with MFA.
  3. Keep one break-glass account under the client's control, stored offline and tested.
  4. Put credentials in a password vault and record who owns each one and when it was handed over.
  5. Never accept passwords by email, chat or a shared spreadsheet.

Physical access belongs on this list too. If the client's server closet, wiring room or office uses keys and fobs, someone needs to know who holds them. A small business access control guide is a good prompt for thinking through who can enter which spaces and how that is revoked when people leave.

Phase 3: Discovery and inventory

Discovery turns the client's description of their environment into evidence. Run network scans, then record IP ranges, VLANs, firewall rules, VPNs, wireless networks and internet circuits. Add cloud tenants, SaaS applications, licenses, and what exists today for backup, antivirus or EDR, and email security.

The inventory is the backbone of everything after it. NIST's small business guidance describes an inventory that records hardware and software identifiers, locations, owners, and the information each technology stores, accesses, processes or transmits. NIST's ransomware guidance also calls for hardware and software inventories and documented information flows as a starting point for managing risk.

A workable inventory has these fields per asset: name, type, location, owner, operating system or version, warranty or renewal date, data it handles, and whether it is covered by monitoring and backup. Gaps you find here become the first remediation list. If you want a structured way to rank those gaps, our SMB IT assessment checklist covers eight areas and how to pick the top three risks.

Good documentation also needs a home that the whole team can reach. If your NOC or service desk works shifts, a network operations center view of the client's circuits, devices and contacts makes after-hours escalations far less painful than hunting through email.

Phase 4: Security baseline

This is the phase generic templates tend to compress into one line. Treat it as its own block, completed before go-live, not after.

Identity first

Change administrative passwords on systems and equipment you inherit. Hunt for user, service and backdoor accounts nobody remembers and remediate them. CISA's Cyber Essentials starter kit recommends MFA for all users where possible, starting with privileged, administrative and remote-access users, together with least-privilege permissions and unique passwords.

Plan MFA like a project rather than a switch. CISA's identity guidance says planning should catalogue user populations, device types and use cases, and account for enrollment, authenticator lifecycle, exceptions and recovery. If the client runs Microsoft 365, our guides on phishing-resistant MFA rollout and conditional access design without lockouts walk through the sequencing.

Connections, patching and backup

  • Review every connection between the client's systems and the provider's. CISA suggests using a dedicated VPN and reviewing those connections, and the same guidance asks customers to check logging, contract terms and incident notification.
  • Review firewall rules and close exposed remote desktop access.
  • Confirm that patches are current, or schedule them with a reboot window the client agreed to.
  • Check that a backup exists, completes, and can actually be restored. A job that reports success is not proof.

For a short list of what to fix first, see our security hardening checklist and the one-page cybersecurity checklist. Keep them separate from onboarding in your head: onboarding sets a floor, and a full assessment goes deeper.

Phase 5: Service setup

Now the tooling goes in, on top of an inventory you trust:

  1. Import network, server, workstation and application data into your documentation and PSA.
  2. Deploy the remote monitoring agent and endpoint protection to every device in the inventory.
  3. Configure role-based user permissions in line with least privilege.
  4. Connect alerts to ticketing, and test that a failed backup or offline server creates a ticket for the right queue.
  5. Set up billing and recurring service schedules, including patching and reporting.
  6. Collect contracts, warranty details and vendor support numbers in one place.

The common miss is the device that exists in the inventory but has no agent. Compare the two lists line by line before moving on.

Phase 6: Kickoff, communication and training

Hold the kickoff meeting once discovery has produced something real to discuss. The agenda is short: introductions, scope and roles, timeline, technical requirements, the SLA, the communication plan, training and open questions. Follow it with a welcome kit.

A useful welcome kit contains:

  • The services in scope and how to request each
  • How to open a ticket, by phone, email or portal, and what each channel is for
  • Escalation paths and after-hours contacts
  • The named person on your team who owns the relationship
  • Milestones for the next 30 days and the date of the first review
  • A short FAQ for end users

Train employees on what changes for them: new MFA prompts, the new way to ask for help, and what a phishing report looks like. Ten minutes of training before go-live prevents a week of confused tickets afterward. Send regular status updates during onboarding even when nothing dramatic happened, because silence reads as drift.

Phase 7: Go-live and hypercare

Go-live is a cutover of the service desk, a welcome message to users, and a short period of closer attention. Before the date, test these five things:

  1. Endpoints report into monitoring.
  2. Alerts fire and land in the correct queue.
  3. A backup restores, not just completes.
  4. Security policies are applied on the devices that matter.
  5. The escalation workflow reaches a human who answers.

Then run a first-week hypercare period: someone watches ticket volume, answers questions quickly and fixes rough edges the day they appear.

Phase 8: The 30-day review

The review separates a finished onboarding from an abandoned one. Reconcile device and user counts against the inventory, review the tickets that came in, confirm backup success and patch compliance, assign owners and dates to remediation items, and revoke the outgoing provider's access. Schedule recurring reporting and quarterly business reviews, then update your template with whatever surprised you.

Revoking the previous provider is the item most often forgotten, and it is a real exposure: an old vendor account with administrator rights is exactly the sort of unused provider access CISA advises disabling.

Building a template you can reuse

A template should have high-level phases and detailed tasks underneath. Give every task an owner, a due date expressed as days from kickoff, and a type: form, document upload, approval, signature, credential handoff, access check or training item.

Do not force every client through one oversized list. Keep a core template for the eight phases above and small add-ons by client type: a medical office adds compliance and device questions, a law firm adds document and matter systems, a fifteen-person office with only cloud apps skips the server section. Last line of the checklist: update the checklist.

Mistakes that cause delays

  • Credentials arrive by email or spreadsheet, or everyone shares one global admin account.
  • MFA is promised but not enforced for administrators.
  • Backups are assumed to work. Nobody ran a restore.
  • Devices without agents sit outside monitoring and patching.
  • Vendor and renewal details are missing, so a certificate or license lapses in month two.
  • Contact and communication plans are vague, and the first outage becomes a debate about who calls whom.
  • The outgoing provider keeps access long after the transition.

Our perspective

We see onboarding as the moment an IT relationship either becomes honest or starts accumulating debt. A small company rarely has a full-time technology leader, so the checklist is often the only place where scope, access, risk and ownership get written down in plain language.

Our own approach starts with a free technology assessment and produces a prioritized plan in plain language that the client owns. That ownership principle applies to onboarding too: the inventory, the documentation and the credential records belong to the business, not to whoever happens to be managing it this year.

We also prefer that anything new, including automation and AI features in the support stack, ships with monitoring, rollback and human checkpoints from the start. An onboarding checklist is the right place to decide who is allowed to do what, because it is much harder to retrofit later.

jaras

Need a second set of eyes on your onboarding?

If you are choosing a provider, rebuilding your own process, or inheriting an environment nobody documented, we can help you find the three risks that matter first. Mindpod Technologies starts with a free technology assessment and turns it into a plan you own. For ongoing guidance without a full-time hire, see our fractional CTO engagements, or start at Mindpod Technologies.

Mindpodtech

FAQ

How long should MSP onboarding take?

It depends on the number of users, sites and how clean the environment is. A small, cloud-based office can be ready in a few weeks, while a multi-site business with legacy servers takes longer. Agree on dates at kickoff and tie them to the phases, not to a single go-live day.

What should I give a new MSP on day one?

Give them your client record, a list of vendors and services, and a way to receive administrator access safely. That means named accounts with MFA and a vault, with a break-glass account you control. Avoid sending passwords by email.

Do I need a different checklist for each client?

Use one core checklist and small add-ons by client type and service scope. That keeps the main list short enough to follow while still covering industry-specific systems like practice management or matter software.

What is the most overlooked item?

Removing the previous provider's access. It is easy to skip once the new service is running, and it leaves privileged accounts open that no one is watching. Put it in the 30-day review with a named owner.

Sources