Skip to content

As of

How long does an ITSM implementation take?

Implementing a service management platform takes months. Keeping it implemented never stops.

Ask how long a new service management platform takes to implement and the answer comes back in months, with a partner, a project plan and a cutover weekend. The question assumes the work ends at go-live.

For most platforms it doesn't. The months go into building a copy of the organization inside the tool: every person, laptop, application and permission, imported and matched. The copy starts going stale the week it launches, and every reorg, acquisition and new application restarts part of the project. The real timeline isn't the months to go live. It's every year you run the platform.

Where the months go

Most of the time goes on two things: the configuration database and the service catalog. Neither is about the software.

The configuration database comes first. The partner imports people from the HR system, accounts from the identity provider and laptops from device management, then matches them. At enterprise scale, thousands of records don't line up on the first pass: contractors with an agency email and a corporate one, people who changed names, laptops still assigned to someone who left. The team fixes them in batches and the import runs again.

Then the service catalog: dozens of request types, each with a form, an approval path and a fulfillment workflow. Before any of it can be built, someone has to write down how each request is handled today and who approves it. Integrations, training and a cutover weekend after a change freeze follow.

By the industry's measure, the implementation is then finished.

Most of the months are spent chasing people

Look closely at those months and little of the time is configuration. Most of it is coordination: finding the person who knows how a request works today, getting them into a meeting, writing down what they said, and waiting for someone else to sign it off.

Every request type in the catalog needs someone who can say how it's handled now, who approves it and which systems it touches. That's a workshop with HR for joiners and leavers, another with finance for purchasing, and one with each application owner for access. Approval paths come from asking managers, and some are out of date by the time the form ships. None of this is hard. It's slow because it runs on calendars across departments.

Implementation workWith a traditional platformWith STLabs
Who works here and who reports to whomImported, matched and cleaned by hand, then imported againRead from HR and identity systems the day they're connected
Devices, apps and who has accessDiscovery projects and reconciliation meetings with each application ownerDiscovered from the integrations and kept current
How each request is handled todayWorkshops and interviews with each department, written up as forms and workflowsLearned from your open tickets, categories and knowledge base
Who approves whatCollected from managers and written into each approval pathTaken from reporting lines in HR, so the current manager approves
Which repeated requests to automateChosen in planning sessions, before anyone sees real volumeSpotted in the requests themselves; the system proposes the workflow
Who qualifies for which accessAgreed by IT, security and the businessStill agreed by people; the agent applies the rules once they're written

The last row is the one no platform removes: deciding policy takes people. Everything above it is either a fact a system already holds or a pattern in work your team already does, and those can be discovered instead of chased.

How long does a service management implementation take?

Published timelines for enterprise implementations commonly run from three months to twelve months or more to go live, with the configuration database alone often taking two to four months.

Where a project lands depends mostly on the organization, not the software: how many systems have to be connected, how clean the identity and HR data is, how many catalog items go into the first release, and how many custom workflows have to be rebuilt. The services work around the license sets the pace, and industry estimates put it at one to three dollars for every dollar of license.

That range answers the question buyers ask. It leaves out what happens next.

Why the implementation never really ends

The configuration database starts to drift the week the platform goes live. People change roles, contractors convert, laptops are swapped, and each change reaches the copy only when the next import runs and someone resolves what it couldn't match. Within a few months someone on the team spends part of every week on reconciliation. It's the import work again, at a smaller scale, with no end date.

Then sales reorganizes into three regions, and dozens of catalog items have approval paths that name the old managers or the old groups. Finding and rewriting them can take a quarter, with the partner back for the workflows nobody wants to touch. Nobody calls it an implementation, but it's the catalog work again.

An acquisition brings a business with its own identity provider, its own HR system and its own service desk. Bringing them in means matching another set of people and laptops into the configuration database, extending the catalog, and deciding whose rules apply. That one gets a project plan, a partner and a target date. It's a second implementation in all but name.

None of this is a failure of the platform or the partner. It's what happens when a tool keeps its own copy of the organization: every time the organization changes, someone has to change the copy.

What an AI agent still can't skip

Two parts take the same time on any platform: the security review of what each integration can read and change, and agreeing who qualifies for which access and who approves the exceptions. An agent can apply those rules once they're written down. It can't write them.

You also choose what history to bring. Open tickets, categories, SLAs and the knowledge base are worth migrating, and Axiom™ reads them so the agent learns how your team solves things. Years of closed tickets only need to stay searchable.

What can go is the copy, and the work of keeping it true.

Implementation with an AI agent and a live model

With STLabs, implementation starts the week the security review signs off.

On day one your team connects Workday, Okta, device management and the service desk you already run. Over the next two days Axiom™ builds a live model of people, devices, apps and access from those systems, joining each contractor's agency and corporate accounts into one identity as it goes. By the end of the fifth day the agent is answering questions in Slack and Teams and resolving the requests you've approved, starting with software access, under rules your team wrote in Policies & Approvals. The old service desk keeps the rest of the queue, in step through bidirectional sync.

There's no configuration database to drift after launch. When someone changes role, the change reaches Axiom™ from Workday as it happens, and the next request uses it.

A reorg arrives as what it is: changes in Workday. New managers, new regions, new teams. Axiom™ already tracks who reports to whom, read from HR, so when a policy calls for the requester's manager to approve, the request goes to the new manager the day the change lands. Nobody spends a quarter hunting through catalog items for the old names.

This is the part of being AI-native that shows up on a calendar rather than in a demo. The agent is only half of it. The other half is that the platform never owned a copy of the organization in the first place, so there is nothing to bring back into line when the organization moves. The work a team used to do after every change is done by the integrations, continuously, and nobody books it as a project.

An acquisition is new connections: the acquired company's identity provider and HR system. Axiom™ joins their people into the same model, including the employees who already had accounts on both sides, and their requests run under the policies already in place. Where the two companies' rules differ, those exceptions are the real work, and the only part that needs people.

How to check a vendor's implementation timeline

Ask two dates, not one.

The first is when a reference customer's system resolved its first real request, counted from contract signature: an access request fulfilled end to end, without an analyst finishing it. A go-live date can mean the portal opened while most of the work was still done by hand.

The second is what happened after that customer's last reorg or acquisition. How long until requests were routing correctly again, and was there a services engagement to get there? Then ask who keeps the platform's picture of people, devices and access current in year two, and how many hours a week it takes.

Start alongside, then consolidate

STLabs runs alongside the service desk you already run, so you can see results on your own requests before you change anything else. Discovery and implementation are included, and teams go from integration to results in five days. What migrates, and what doesn't need to, is in Moving to STLabs from your current service desk.

The useful question was never how long the implementation takes. It's how long until you stop doing it.

Frequently asked questions

How long does an ITSM implementation take?

A traditional implementation commonly takes three to twelve months to go live, and parts of it repeat after every reorg or acquisition. STLabs goes from integration to results in five days, with no copy of the organization to rebuild.

Why do ITSM implementations take so long?

Mostly because the tool needs its own copy of the organization: people, devices, applications and access imported, matched and cleaned, plus a catalog and workflows configured one by one.

What happens to an ITSM platform after a reorg or acquisition?

Approval paths, catalog items and configuration records that reference the old structure need reworking. A platform that reads roles and reporting lines live from HR picks up the new structure as it changes.

Can we run a new ITSM platform alongside the old one?

Yes. With two-way sync, the new platform can take on one request type at a time while the old tool keeps the rest of the queue, so there's no single cutover date.

See it on your tools.