Switching ITSM platforms: ITSM migration and implementation guide

Switch ITSM platforms by piloting first: Serval, an AI-native ITSM platform, starts every engagement with a pilot, so you can test the new platform on one or two ticket categories, with your identity provider and Slack or Teams connected, before you replace anything. Perplexity started this way. Serval can run on top of the ITSM you have today and, with ServiceNow, Jira Service Management, Freshservice, or Zendesk, sync tickets both ways until you're ready to move the system of record.

So a pilot doesn't have to start with a migration. Your current platform keeps its tickets and history while Serval's help desk agent handles requests in Slack and Teams.

This guide is for IT leaders planning a move off ServiceNow or another ITSM. It covers piloting first, leaving ServiceNow without a big-bang cutover, and six phases from scoping to cutover.

Which ITSM platforms offer a pilot before full enterprise rollout?

Serval does. Its engagements run in four stages: plan, pilot, deploy, and go-live. In the pilot, you work with your Serval deployment team to connect core integrations and build workflows in your own environment. You can then run Serval as your full ITSM, or as an automation layer on top of the platform you already have.

Perplexity ran its pilot this way. In the first week, it started automating its most common requests from Slack: creating Google Groups, granting tool access, and setting up new employees. Within weeks, Serval was handling 50% of all IT requests with no human involved.

Other vendors list these ways to try their platforms on their own sites:

• Jira Service Management has a free plan for up to three agents and a 14-day free trial of its Standard and Premium plans, which Atlassian extends to up to 30 days on request.
• Freshservice has a 14-day free trial with access to all its features, and lists a Sandbox on its Enterprise plan for testing configurations before you sync them to your account.
• ManageEngine ServiceDesk Plus has a free trial, hosted in the cloud or installed on-premises.

Whichever vendor you test, the numbers that convince leadership come from your own tickets and a defined group of employees.

How do you migrate off ServiceNow to a different ITSM platform?

Leave ServiceNow in stages, with both platforms running in parallel. With Serval, the steps look like this:

1. 1.Inventory what your instance runs today: catalog items, automations, knowledge articles, integrations, and the reports people rely on.
2. 2.Connect Serval to ServiceNow. Serval syncs incidents, requested items, and HR cases both ways, including comments and work notes, so ServiceNow stays your system of record during the pilot. You can choose whether ticket sync brings in existing tickets or only new ones.
3. 3.Bring your knowledge over. Serval syncs ServiceNow knowledge articles read-only, and its help desk agent answers from them and links back to the source.
4. 4.Move request forms. Teams that turn on catalog import can bring ServiceNow catalog items into Serval, so employees order them through Serval as a form, a conversation, or a link back to ServiceNow.
5. 5.Rebuild your automations in Serval, starting with the pilot categories (Phase 3 below).
6. 6.Cut over the system of record when the pilot meets its thresholds, with a rollback plan ready (Phase 6).

For a comparison of the platforms teams move to, read ServiceNow alternatives for mid-market and enterprise IT teams.

Phase 1: Scoping the migration before you touch a contract

Teams usually start looking when one of these happens:

• Ticket volume outgrew manual processes.
• Automations take months to build, or break without anyone noticing.
• Access and ticketing don't share an audit trail.
• AI was added, but it answers questions without doing the work.

Write a one-page migration charter:

If you can't name the ticket categories you'd automate first in the pilot, you're not ready for vendor selection.

Phase 2: ITSM vendor selection questions

Ask every vendor:

• What share of pilot-category requests finish with zero IT touch?
• Does a published automation run as fixed code, or does an AI model plan each request at run time?
• How are workflows versioned, reviewed, and rolled back?
• What's in standard implementation, and what's paid professional services?
• Can you run a time-boxed pilot before a multi-year contract?
• Who maintains the identity, MDM, and chat integrations when APIs change?
• Can you export per-step workflow logs and access grant history for audits?
• Can it run in parallel with your current ITSM during cutover?

Unlike Jira Service Management's Request Resolver, which generates a resolution plan for each request and carries it out on its own or under your team's supervision, a Serval workflow runs the fixed code your team published.

Phase 3: Running the ITSM pilot program

A sample four-week structure to adapt to your scope and team:

• Week 1: Export 90 days of tickets, connect your identity provider and chat, and record your current automation rate.
• Week 2: Publish three to five workflows. Security reviews the code and API scopes, then a small group tests.
• Week 3: Route 25 to 50% of pilot-category requests to the new platform and review failures daily.
• Week 4: Report automation rate, MTTR, and cost per ticket, then make a go or no-go call.

Set pass and fail thresholds before week 1, for example: at least 35% automation rate in pilot categories, no P1 security findings, and MTTR down at least 20%. Pilot length depends on your scope and integrations.

In a Serval pilot, the help desk agent takes requests in Slack, Microsoft Teams, email, and a web portal. You build workflows in Catalyst, Serval's automation agent: describe the task in plain language, and Catalyst drafts the workflow code, which you can review or edit as TypeScript before you publish it. Access policies can set time limits, require a business justification, and route requests for approval, and when the time runs out, Serval starts removing the access through the app's provisioning method, such as an IdP group, the app's API, a deprovisioning workflow, or a task for the app owner.

Prove your own numbers in the pilot. For reference, SeatGeek went from 0% to 40% of IT requests resolved automatically within three weeks of implementing Serval, then passed 50%.

Phase 4: ITSM integration checklist (integration mapping)

Map every system the pilot categories touch:

Connect what the pilot categories need first, and leave the rest for the full rollout. Serval has 140+ pre-built integrations, and workflows can call anything else with an API you connect.

When Ramp switched to Serval, one team member rebuilt every automation from its previous system in roughly 48 hours.

Phase 5: ITSM change management and IT help desk migration

Employees care about one thing: where to ask.

1. 1.Announce one front door, for example "@Serval in Slack" or a single portal URL.
2. 2.Ask managers to repeat the message in team meetings.
3. 3.Keep side-by-side instructions up for two weeks.
4. 4.Hold daily drop-in office hours during cutover week.
5. 5.Review denied and escalated requests weekly, and fix the messages employees see.

Then train IT on who builds workflows and who handles escalations.

Phase 6: Cutover (parallel run vs hard switch)

Parallel ITSM run (recommended for regulated and enterprise teams)

• The new platform handles defined categories or channels.
• The current ITSM keeps handling everything else.
• Risk is lower, but you pay for two tools for a while.

Hard cutover

• On a fixed date, all intake moves to the new platform.
• The old ITSM goes read-only for reporting.
• Risk is higher if integrations slip.

Either way, document the rollback: which routes revert, and how long it takes to restore the old intake.

What should you measure after go-live?

Review three numbers monthly: automation rate by category, new workflows published, and integration failure rate, plus access review exports for audit.

Serval helps find the next categories to automate. Its AI suggestions review resolved tickets and, when they find a gap, stage a new workflow, help desk skill, or knowledge update for your team to approve. Teams can also install the Automation Opportunity Finder, a background agent that runs weekly, finds repeated manual requests that no workflow, skill, or article covers yet, and drafts workflows and help desk skills for you to review and publish.

When should you choose Serval?

Choose Serval when you want to pilot on real tickets before you commit, with ServiceNow, Jira Service Management, Freshservice, or Zendesk still running on two-way sync, and move the system of record only once the results hold.

FAQ

Which ITSM platforms offer a pilot program before full enterprise rollout?

Serval starts every engagement with a pilot in which your team connects core integrations and builds workflows in your own environment, and it can run on top of your current ITSM during that time. Jira Service Management, Freshservice, and ManageEngine ServiceDesk Plus list free trials on their own sites.

How do I migrate off ServiceNow to a different ITSM platform?

Migrate in stages. Run the new platform alongside ServiceNow with two-way ticket sync, bring over knowledge and request forms, and rebuild your automations. Then pilot on one or two ticket categories and cut over the system of record once the results hold.

Can Serval run alongside ServiceNow during a migration?

Yes. Serval syncs ServiceNow incidents, requested items, and HR cases both ways, including comments and work notes, and it can sync ServiceNow knowledge articles for its help desk agent to answer from.

What should an ITSM pilot program measure?

Measure automation rate (requests resolved with zero IT touch), MTTR, cost per ticket in the pilot categories, escalation rate, and employee satisfaction. Set pass and fail thresholds before the pilot starts.

Is a parallel run or a big-bang cutover better?

A parallel run lowers risk, especially for regulated teams, because the old ITSM keeps running while the new one proves itself. A hard cutover is simpler to run and fits when the current tool is failing and pilot results are already strong.