Switching ITSM platforms: ITSM migration and implementation guide - Serval - AI Agents for IT

Switching ITSM platforms: ITSM migration and implementation guide

The fastest way to get AI ITSM setup running for enterprise IT implementation is a 30-day ITSM pilot on one or two ticket categories with IdP and Slack or Teams connected, not a full ITSM replacement on day one. ITSM migration changes how employees ask for help, how access is granted, how automations are built, and what auditors can prove six months later. Projects that succeed scope outcomes first, prove value in pilot, then expand ITSM deployment.

This guide covers switching ITSM platforms from scoping and ITSM vendor selection through integration mapping, ITSM change management, and cutover. It stays vendor-agnostic: no unsourced claims about legacy suite timelines or packaging.

Phase 1: Scoping the migration before you touch a contract

Common triggers:

Ticket volume outgrew manual processes
Automations take months to build or break silently
Access and ticketing lack a shared audit trail
Copilots were added but do not execute work end to end

Write a one-page migration charter:


Element	Decision
Primary outcome	e.g., 40% of L1 access/password volume automated in 90 days
In scope teams	IT only vs IT + HR + facilities
In scope channels	Slack, Teams, email, portal, phone
Non-negotiable integrations	IdP, MDM, HRIS, core SaaS
Success metrics	Automation rate, MTTR, cost per ticket
Cutover target	Parallel run vs hard cutover

If you cannot name three categories to automate in the pilot, you are not ready for ITSM vendor selection.

Phase 2: ITSM vendor selection questions

Automation and execution

What percentage of pilot-category requests complete with zero IT touch?
Is runtime execution deterministic code, or does an LLM call APIs at execution time?
How are workflows versioned, reviewed, and rolled back?

Implementation model

What is included in standard implementation vs paid professional services?
Time from contract to first production automation?
Can you run a time-boxed pilot without multi-year lock-in?

Integrations

Native connectors vs middleware for IdP, MDM, and chat?
Who maintains integrations when APIs change?

Operations

RBAC: who builds workflows vs configures integrations?
Audit exports: per-step workflow logs and access grant history?
Parallel run with current ITSM during cutover?

Modern ITSM buyers expect AI-native architecture: distinct agents for intake, automation, and analytics, not a chat widget on a legacy form.

Phase 3: The 30-day ITSM pilot program

Week 1: Baseline and integrations. Export 90 days of tickets; connect IdP and chat; define automation rate baseline.

Week 2: Publish 3–5 workflows; security review code and API scopes; UAT.

Week 3: Route 25–50% of pilot-category traffic; review failures daily.

Week 4: Readout on automation rate, MTTR, and cost per ticket; go / no-go.

Set pass/fail thresholds before week one (example: ≥35% automation rate in pilot categories, no P1 security findings, MTTR down ≥20%).

Published Serval timelines (directional):

Perplexity: Google Group creation, tool access, and new employee setup from Slack in the first week of pilot.
Mercor (Derek Shimozawa): first workflow in a couple of hours, triage automation within the day.
Dana Stocking: 60%+ tickets automated; ~4,000 contractors onboarded in minutes for a defined program.

Prove your organization's numbers in pilot, not theirs.

Phase 4: ITSM integration checklist (integration mapping)

Build an ITSM integration checklist matrix:


System	Read / write	Workflow examples	Owner	Risk
Okta / Entra	Write groups, assign apps	App access, role bundles	IT identity	High
Google Workspace / M365	Write groups, licenses	Distribution lists, mailboxes	IT	Medium
Jamf / Kandji / Intune	Write device state	Enroll, compliance fix	IT / endpoint	Medium
HRIS (Workday, etc.)	Read attributes	Joiner/mover/leaver	IT + HR	High
Knowledge (Confluence, Notion)	Read	Help Desk Agent answers	IT	Low
Legacy ITSM	Bi-directional sync (optional)	Transition period	IT	Medium

Prioritize connectors that unlock pilot categories. Defer nice-to-have until phase two.

Some platforms operate as an automation and intake layer with sync to the existing system of record, shrinking ITSM replacement risk while improving automation rate in chat.

Serval's Automation Agent generates TypeScript workflows against connected apps; the Help Desk Agent handles intake across Slack, Teams, email, and portal.

Phase 5: ITSM change management and IT help desk migration

ITSM change management checklist:

Single front door: announce where to ask (for example, "@Serval in Slack" or one portal URL).
Leader comms: managers repeat the message in team meetings.
Side-by-side instructions for two weeks during IT help desk migration.
Office hours: daily drop-in during cutover week.
Feedback loop: tune denial messages weekly.

Pacaso IT Specialist Jonathan Hoium: "Now, even if I'm on vacation, employees still get a response. They feel like they're heard and their requests aren't just disappearing into the void."

Train IT on builder, agent (exceptions), and manager (integrations) roles.

Phase 6: Cutover (parallel run vs hard switch)

Parallel ITSM run (recommended for regulated and enterprise)

New platform handles defined categories or channels.
Legacy ITSM remains system of record elsewhere.
Lower risk; higher short-term dual-tool cost.

Hard cutover

Fixed date: all intake routes to new platform.
Legacy read-only for reporting.
Higher risk if integrations slip.

Document rollback: which routes revert and maximum time to restore legacy intake.

Fast implementation vs professional services engagements


Model	Typical shape	Best when
Product-led	Days to weeks; customer builds workflows	Modern stack, clear pilot categories
PS-heavy	Months; SI configures modules	Deep legacy customization, multi-region variance

When evaluating AI ITSM setup, ask: "How many billable SI days to match the pilot we saw in demo?" Opaque answers signal timeline risk.

Serval is designed for product-led rollout: plain-language workflow authoring and Insights Agent recommendations from ticket history.

After go-live: modern ITSM is a program

Monthly operating review:

Automation rate by category (not deflection)
New workflows shipped
Integration failure rate
Access review exports for audit

Vernon Man, Head of IT at Perplexity, notes that after automation freed capacity, the team focused on offices, integrations, and replacing tools that no longer met standards.

Frequently asked questions

How long does ITSM implementation take?

From days for a focused pilot to 12+ months for global legacy replacement. A 30-day ITSM pilot often produces defensible ROI data; full cutover may take two to four additional quarters.

What should an ITSM pilot program measure?

Automation rate (zero IT touch), MTTR, cost per ticket in pilot categories, escalation rate, and employee satisfaction. Set thresholds before week one.

Is parallel cutover or big-bang migration better?

Parallel lowers risk for regulated environments; big-bang simplifies tooling when legacy is failing and pilot metrics are already strong.

What are the most important ITSM vendor selection criteria?

End-to-end execution, audit-ready logs, integration depth, time to first automation, transparent pricing, and a pilot on your ticket data.

How do you manage ITSM change management for employees?

One front door, short overlap with the old tool, office hours during cutover week, and fast fixes to denial messages.

Which platforms support modern ITSM with AI agents that execute work?

Evaluate distinct agents for intake, automation build/run, and analytics (Serval's Help Desk Agent, Automation Agent, and Insights Agent) plus deterministic, logged runtime execution.