Skip to main content
View as MarkdownOpen in ChatGPTOpen in Claude

Identity Mesh — May 20, 2026

Release Date: May 20, 2026
Release Type: Mixed (New Features, Improvements, Bug Fixes)

Overview​

This release introduces conversational tenant onboarding — a guided, chat-based setup experience for administrators signing in to a new tenant for the first time.

The two known issues from the May 13, 2026 release — the risk score tooltip showing only a generic label, and the asset details panel hiding content for crypto assets — are fixed in this release. A data-integrity correction also closes a long-standing gap where partial software inventory updates could overwrite complete machine asset records.

This release also includes reliability and supportability improvements.

New features​

Conversational tenant onboarding​

When an administrator signs in to a tenant that hasn't completed setup, they're now guided through a full-page conversational onboarding experience instead of landing on the dashboard. The onboarding assistant walks you through the whole setup in one continuous flow:

  • Capture your organization profile and role.

  • Detect integrations that are already configured — for example a pre-provisioned Microsoft Entra ID connection — and acknowledge and skip them.

  • Guide identity provider setup and recommend connectors for your environment.

  • Validate each configured integration and run its first sync.

  • Finish with a summary of what was configured and how the first syncs went.

How the flow behaves:

  • Only capabilities enabled for your tenant are offered; unavailable options are never shown.

  • Progress is saved at the tenant level and shared across all administrators, so anyone can resume exactly where the flow left off — even after an interruption.

  • If you skip onboarding, you return to the flow on your next sign-in. Users without administrator access go straight to the dashboard, with a note that onboarding is incomplete.

  • If several administrators onboard the same tenant at once, both sessions reflect the latest saved progress.

  • If validation or the first sync fails for an integration, onboarding stays resumable, with that item marked incomplete and the remaining work summarized.

When everything completes, onboarding is marked done and you're taken to your dashboard. This replaces the manual setup that previously required moving between several admin pages, and shortens the time from first sign-in to a working tenant.

Improvements​

Machine asset records no longer lose fields after software inventory updates​

Software inventory updates carry only a subset of a machine asset's fields. Previously, when a partial update arrived after a complete one, it could replace the full record and silently erase field values. Partial updates are now merged into the existing record instead, so machine asset fields are preserved across the full update lifecycle and no longer disappear after a software inventory update.

Same risk score tooltip on the asset list and the asset details page​

The risk score tooltip on the asset list didn't match the one on the asset details page — the same score could appear to mean different things depending on the view. Both views now show identical tooltip content and behave the same way:

  • The same text: a header, one line per direct risk on the asset, and a summary of risks inherited from correlated assets.

  • The same behavior for when and how the tooltip appears and dismisses.

  • Both views draw from a single source, so they stay in step automatically in future updates.

Risk score calculation and the visual score display are unchanged, and the existing behavior when a score is unavailable is preserved.

Bug fixes​

Workday connector failures now visible in connector health​

When data collection from Workday failed, the connector kept retrying the same failing batch without reporting the error — so a stuck sync looked healthy and only showed up as missing data downstream. Failures are now recorded, and the reason appears in the connector health view, so you can see why a Workday sync is stuck. Retry behavior is unchanged.

Detailed risk score tooltip restored on the asset details page​

Fixes a regression listed as a known issue in the May 13, 2026 release notes, where the risk score tooltip on a machine asset showed only a generic severity label (for example "Critical Risk") instead of explaining why the asset has its score.

Pointing to the risk score on the asset details page now shows the full breakdown again:

  • A header: "The asset's score reflects:"

  • One line per direct risk category on the asset, with its count — for example "Risks associated to this asset: No owner (3)"

  • A summary of inherited risks — for example "Risks inherited from 38 correlated assets (190)"

Severity is still conveyed by the color of the score. This restores the at-a-glance answer to "why is this asset risky?" during triage, without navigating to other views.

Asset details panel scrolls again for crypto assets​

Fixes a regression listed as a known issue in the May 13, 2026 release notes, where the left panel on a crypto asset's details page had no way to scroll. Content below the visible area was cut off, hiding the Risks section and the "Delegate investigation" and "View in ServiceNow" buttons.

The panel now scrolls whenever its content is taller than the window, so all fields, the Risks section, and both action buttons are reachable on standard screen sizes. On windows tall enough to fit everything, no scroll bar appears. You no longer need to zoom out or resize the browser to reach these controls.

Version, Product, and OS version columns now sortable​

On the software asset list, the Version and Product columns didn't respond to clicks and couldn't be sorted, while the other columns sorted correctly. The OS version column on the machine asset list had the same issue. All three columns now sort correctly, with the same ascending and descending indicators as the other columns.

Dashboard exports can be retried after a temporary error​

If a temporary service error occurred while starting an export from the dashboard, the export showed a generic "Export failed" message — but trying again produced "A similar export job is already in progress," leaving no way to retry or cancel. You had to wait for the invisible job to finish or expire.

What changed:

  • After a temporary error, the export now enters a verifying state and checks the actual job status instead of declaring failure.

  • While an export is running, the export button is disabled and the job's progress is shown.

  • If an export is already running, its status is displayed instead of an error message.

There's no longer any ambiguity between "export failed" and "export running," and you're never left without a way to retry.