Microsoft retired InfoPath on 14 July 2026. The records still have to be preserved — this checklist gets it done cleanly.

Use this as a working checklist. It’s ordered so your originals are never at risk, and so that at the end you have not just the data but the evidence that you preserved it properly.

1. Scope & inventory

  • Locate every InfoPath .xsn template across PCs, shares and SharePoint.
  • Locate every filled-in form-data .xml record set.
  • Map each to its business process and its records-retention class.
  • Note approximate record counts and any known legal holds.
  • Flag which forms are historical records (archive) vs still-active workflows (migration).

2. Protect the originals

  • Copy all sources into a read-only “originals — do not modify” store.
  • Confirm your export method never writes back to a source file.
  • Record a simple integrity check (e.g. file counts and sizes, or hashes) of the originals.

3. Recover the data

  • Extract the field catalogue from each .xsn (from its .xsd).
  • Recover every filled-in value from the form-data .xml, including repeating/nested sections.
  • Export to open formats — CSV and JSON at minimum — that will outlive any one application.
  • Keep a per-file record of what was recovered and anything that wasn’t.

XSN Rescue produces the last point automatically: an honest recovery report on every run stating exactly what was and wasn’t recovered, integrity notes, and a source_modified: false read-only record per file. If you recover by hand, keep the equivalent notes yourself — auditors value the honest record of gaps as much as the data.

4. Build the archive

  • Create a durable, human-readable copy of each record (PDF or HTML) that opens without special software.
  • Produce an index so any record can be located quickly.
  • Produce a summary/inventory of the whole set for sign-off.
  • Preserve the original on-screen layout where it matters (the .xsl views), noting where complex views render only partially.

If assembling that by hand is daunting at volume, XSN Rescue’s archive output generates it from one local pass: an archival PDF per record, an archive-manifest.csv plus a browsable archive-index.html, and an estate-inventory executive report (HTML, JSON and a one-page PDF) — nothing uploaded.

5. Store, retain & document

  • Keep the readable archive alongside the untouched originals.
  • Store a copy in a second location per your backup and retention policy.
  • Record who performed the archival, when, with what tool/version, and the outcome per source.
  • Diarise a future readability check (e.g. can we still open a sample record in N years?).

6. Handle the “not archivable” honestly

  • Note that form code-behind, external data connections and user-role logic are application behaviours, not archivable records — scope a migration separately if those workflows are still active.
  • Note that digital signatures are preserved as recorded values, not re-verified cryptographically.
  • Note that decodable attachments come back as real files, and that any embedded content that could not be decoded is preserved and marked rather than guessed.

Comfort & control, the whole way through. The point of this checklist is that when it’s done you can say, truthfully: nothing was lost, nothing original was touched, and everything we kept can be opened and read. That’s a good place to be standing now.