Skip to main content
View as MarkdownOpen in ChatGPTOpen in Claude

Identity Mesh — August 3, 2026

Release Date: August 3, 2026
Release Type: Mixed (New Features, Improvements, Bug Fixes)

Overview​

This release introduces a new correlation engine, two new crypto-posture widgets, and a new way to explore identity relationships:

  • MeshIQ (AI-based correlation engine) — supersedes the previous correlation engine. You choose when to move, and you can always see which model your organization is on.

  • Quantum-vulnerable assets and compliance status widgets — two new widgets covering post-quantum exposure and policy compliance, each with drill-down to the assets behind the number.

  • 3D identity and asset graph — a new interactive view for following how identities and assets relate.

It also moves onboarding and security-code emails onto the Axiad Mesh template, restores per-widget chat history in Generative UI workspaces, and fixes five problems in asset filtering, widget totals, and risk score display.

New features​

MeshIQ (AI-based correlation engine)​

Mesh correlates accounts from your connected systems into single identities. That matching runs on a model, and this release introduces MeshIQ — an AI-based matching model that supersedes the correlation engine used until now. MeshIQ matches with strict per-organization isolation: your data is never matched against anyone else's.

What changed:

  • Your organization has exactly one active model at a time. The Correlation Engine card under System configurations → General shows which one — for example, MeshIQ · 1.

  • Models are versioned. Axiad publishes new versions as they are released and makes them available to your organization, but you are never moved onto one automatically.

  • Re-Correlate / Change Model moves your organization onto a model you select and re-matches your existing identities against it. You choose the timing, and the card shows progress while it runs.

  • Organizations can request a model trained on their own data. This is being rolled out gradually and is enabled separately — if you do not see the Request Tenant-Specific Training button, ask Axiad Support whether it is available for you.

Why it matters: matching quality decides whether Mesh shows you one person or five disconnected accounts. Putting the model on a version you control means you can adopt an improvement deliberately, and know exactly what your identities were matched with.

See Manage the correlation engine for how to check your active model, re-correlate, and request tenant-specific training.

Quantum-vulnerable assets and compliance status widgets​

Two new widgets are available for your workspaces and dashboards.

What you get:

  • Quantum-vulnerable assets classifies your certificates and keys by post-quantum exposure, so you can see how much of your estate relies on algorithms that a quantum computer would break.

  • Compliance status shows your assets against the policies in force for your organization.

  • Both support drill-down: selecting a bucket opens the matching asset list, filtered to that bucket, so you can act on the assets behind a number.

Why it matters: post-quantum readiness and policy compliance are things you report on, not just observe. Both widgets resolve to a concrete asset list, so a number in a board pack traces back to the assets that produced it.

3D identity and asset graph​

A new interactive graph renders identities and their related assets in three dimensions, so you can follow relationships by moving through them instead of reading them out of a table.

What you get:

  • The graph is available on identity and asset surfaces.

  • It uses the same interaction model everywhere it appears, so what you learn on one surface carries to the others.

Why it matters: blast radius and shared-ownership questions are relationship questions. Seeing the shape of the graph answers them faster than paging through a list.

Improvements​

Per-widget chat history restored on revisit​

In Generative UI workspaces, the chat you had with an individual widget was lost when you navigated away and came back. That history now persists, so returning to a workspace picks up where you left off.

Widget chat also gained clearer grounding in the underlying data, and more visibility into the arguments a query ran with.

Better answers for asset questions​

Asking about assets in a workspace now returns more consistent results, particularly where the question is scoped to cryptography or asks for a ranked list. Ranked answers come back as a table rather than prose.

Onboarding and security-code emails​

The emails Mesh sends for onboarding links and security codes now use the Axiad Mesh template, so they match the rest of the product rather than the previous generic styling.

Smaller improvements​

  • The account menu now reads Mesh Docs and Mesh Status, pointing at the current documentation and status URLs.

  • Signature algorithm names are normalized for display, so the same algorithm no longer appears twice under different spellings.

Bug fixes​

Top 10 asset widgets showed an unstable set of rows​

The Top 10 asset widgets could return a different set of rows between refreshes when several assets tied on score. Ordering is now deterministic, so repeated views of the same data show the same assets in the same order.

Asset filters returned no results for some fields​

Certain asset filters were sent with field names in the wrong casing and silently matched nothing. The field names are now normalized, and the affected filters return the expected assets.

"Machines 0" shown in estimated assets in scope​

The estimated assets panel displayed a Machines 0 cell even where no machine estimate applied, which read as "you have no machines" rather than "not applicable". The cell no longer appears in that case.

Risk score briefly blank after a scoring run​

An organization's risk score could render blank while a scoring run was in progress. The last known value is now carried forward until the new one is available.

Algorithms flagged twice in post-quantum grading​

An algorithm that had already been graded could also be counted as having insufficient parameters, so the same asset appeared under two findings. Graded algorithms are no longer double-flagged.

Deprecations​

Previous correlation engine​

The correlation engine used before this release is superseded by MeshIQ and is being retired.

  • All new ingestion is already scored by the newer engine — new accounts and new data on existing accounts alike. Your existing correlations stay as the previous engine produced them until you re-correlate.

  • Every organization moves onto a MeshIQ model eventually. Use Re-Correlate / Change Model to move at a time you choose — see Manage the correlation engine.

If you have not been offered a MeshIQ model, contact Axiad Support.