I work at Help Desk Migration, which builds migration apps for Jira Service Management on the Atlassian Marketplace. The reconciliation approach below applies to whatever tool or script you use.
For most teams, a service desk migration ends when the dashboard says “completed,” and agents start working in the new JSM project.
For a regulated company, that’s where a different question starts. Months later, an auditor, a security reviewer, or your own compliance team will ask: how do you know every record made it across?
“The tool said it finished” doesn’t answer that. Neither does “we checked a few tickets”. This post explains why, and what a defensible answer looks like.
Why tickets are records, not just history
In a regulated environment, a service desk is not only a support tool. It’s where the evidence lives:
- Access requests and approvals. Who asked for access to which system, who approved it, and when.
- Incidents. When an issue was reported, who handled it, and how it was resolved.
- Change and onboarding records. Proof that a defined process was followed.
- Customer and employee requests that you’re obliged to keep for a set period.
When that history moves from one platform to another, the migration itself becomes part of the audit trail. The question isn’t only “did the data move?”. It’s “can you show it moved completely, and who confirmed it?”
What reviewers actually ask after a platform change
From the migrations we support, the questions after a regulated team switches platforms tend to be the same:
- Completeness. Is every record in the target that was in the source? If not, why not?
- Accuracy. Did values, relationships, authors, and timestamps survive, or were they changed along the way?
- Explained differences. Every gap needs a documented reason: filtered on purpose, matched to an existing record, or failed and re-migrated.
- Sign-off. Who confirmed the result, when, and based on what evidence?
- Retirement. Was the old system switched off only after the new one was verified?
Spot-checks don’t answer these at scale. A spreadsheet of before-and-after counts answers the first one, partly. A reconciliation report covers completeness and traceability, and gives you the baseline for the rest. Accuracy of values still needs a human check.
Why counts alone mislead in Jira Service Management
Total counts are where most teams stop, and in JSM they often look wrong when the data is right:
- Customers are matched, not duplicated. If the same person appears twice in the source, JSM gets one customer.
- Some fields have no native home. Source groups, CC users, or original reporters may go into custom fields, tags, or the description. A native JSM report won’t find them there.
- Attachments can go up. Inline images migrated as attachments raise the total.
- Records created during the run only arrive with the delta migration.
Each of these is a legitimate difference. Without a documented explanation, each one also looks like data loss to a reviewer. That’s the gap a reconciliation report helps close.
A real example: Vanta’s move to Jira Service Management
Vanta, the compliance automation company, moved its internal service desk from Freshservice to Jira Service Management. The scope was 128,000+ tickets, 512,000 comments, 86,000 attachments, and 4,200+ requesters and agents. JSM is now the request portal for IT, People Ops, Workplace, GTM Systems, HR, and Access Management.
Janiece Caldwell, who leads IT Operations in Vanta’s Corporate Engineering team, explained why the history mattered:
“We’re a compliance company, so being able to show timestamps and request history matters for things like our own SOC 2 audit.”
So every ticket had to keep who requested it, who handled it, who approved it, and when. Freshservice and JSM fields don’t line up one-to-one, and JSM had no direct equivalent for the original reporter and assignee on historical tickets. The migration was mapped around that:
- The original reporter and assignee went into custom Original reporter and Original assignee fields. The native fields were left as Anonymous and Unassigned.
- Approvals moved into the ticket description, with the approver’s name, email address, and approval date.
- Agent comments start with the original author’s name and email.
- Freshservice groups became tags with an FS- prefix. Inline images became attachments.
Every one of those decisions is correct. Every one also means native JSM reports won’t match the source: requests show “Anonymous” as the reporter, groups appear as tags, and the attachment total is higher. For a compliance team, that’s exactly why a matching ticket count is not enough. You need a record of where each piece of context ended up, and a check that it’s all there.
Vanta also validated the mapping on sample data before the full run. That surfaced the reporter and assignee issue early, so it was fixed before the real migration instead of being found later. The final transfer was a re-migration, run after the customizations and testing were complete.
What the Data Reconciliation report gives you
This is why we built the Data Reconciliation report into Help Desk Migration. After a migration, you download an archive with an HTML report and a JSONL file. For a regulated team, its value comes down to five things:
- A documented record of completeness. The Migrated Entities table shows, per entity type, how many records were available in the source, migrated, skipped, failed, or not selected. It replaces “we think everything moved” with numbers.
- Record-level traceability. The Entity Migration Report maps every source ID to its target ID. For any skipped or failed record, it shows the status and the reason. Load records.jsonl into the report and the data is processed in your browser, not sent to a server.
- Evidence you can hand over. Teams use the report for internal data governance, compliance reviews, and stakeholder sign-offs.
- Your setup on file. Applied filters and customizations are listed in the Summary. The extended report also saves the full Field Mapping Summary (which source field went to which target field, including value mappings and defaults), the migration options, and the migration history with timestamps.
- A clear baseline for sign-off. You can retire the old system once the numbers reconcile and your spot-checks are done, not on a feeling.
A few things to know before relying on it:
- There are two versions. The default report is free and available at the Demo or Full step. The extended report is a paid add-on at the Full step (the Full reconciliation report option). Field mapping, options, history, and access details are in the extended version only.
- It doesn’t replace judgment. The report tracks records and statuses, not field values. You still spot-check context: approvals, authors, and timestamps.
- It doesn’t record your sign-off. Migration History shows status changes and Migration Access lists who can open the migration, but who confirmed the result, and when, you record yourself.
- Not every gap is a skip. “Skipped” and “failed” records come with a reason in the report. Records excluded by a filter, or customers matched to an existing one, are differences you should document yourself.
- A Migration Score below 100 is normal. The score (1 to 100) shows how much of the source structure you included in the migration. It doesn’t measure record-level accuracy. A score like 81 is labeled “Partial coverage” and usually means you skipped some entities or options, or the platforms are structured differently.
A short sign-off checklist for regulated teams
- Take a per-record-type baseline in the source before migrating.
- Document every filter and mapping decision, especially where data moves into custom fields, tags, or the description. Keep the extended report’s mapping summary if you have it.
- Reconcile after the full run, per record type, and re-check the counts after any delta run.
- Resolve failed and skipped records, and record the reason for anything else left out (filters, matched customers).
- Spot-check context: approvals, original reporters, comment authors, and timestamps.
- Record who signed off and when, and store the reconciliation output with that sign-off, for as long as you keep the records.
If you’d rather not build the comparison yourself, our apps are on the Marketplace: Help Desk Migration on Atlassian Marketplace.
If you work in a regulated industry, what did your auditors ask for after your last platform change? I’d like to compare notes.