October 30, 2026 SaaS Alerts release notes

Ask AI - MSFT Agent

Enhancements

Compliance framework scores and gaps are available

Ask AI can now report how an organization scores against compliance frameworks, including HIPAA, CIS Controls, and NIST SP 800-171, using Fortify's latest assessment.

Results include the percentage of controls met, the number of controls that are Met, Partially met, Not met, Not assessed, or require Manual review, and the date of the assessment.

Ask AI can also identify controls that are not currently met and the Fortify actions that can address them, with the quickest wins presented first.

Framework results can be compared across organizations to identify organizations that are furthest from meeting a framework.

Compliance scores cover Microsoft 365 settings only and do not represent certification.

Compliance assessments can be run from Ask AI

Ask AI can now start a new compliance assessment against a framework for one or multiple organizations.

After starting an assessment, Ask AI identifies the assessments that were started and displays each organization's previous result while the new assessments are completed. You can then request the results to see the updated score, counts for each assessment status, and controls that are still not met.

Ask AI also offers to run a new assessment when an organization has not previously been assessed, when its latest assessment is more than one week old, or when the assessment was completed before actions were applied.

Compliance framework answers use Fortify's reviewed mappings

Ask AI now uses Fortify's reviewed compliance mappings when answering questions about supported compliance frameworks, organizational framework requirements, and templates aligned with a framework.

Responses identify the framework version and the number of framework controls covered by Fortify actions. They also identify the controls associated with each action and whether the action covers each control fully or partially.

For questions about a specific organization, Ask AI also reports the status of the applicable Fortify actions.

When Fortify does not have a mapping for a framework, Ask AI identifies its recommendations as a best fit rather than presenting them as mapped Fortify guidance.

Fixes

Secure Score improvement values are more accurate

Ask AI no longer reports the full value of a recommended action as the potential Secure Score gain when the action has already earned some points.

It now reports only the additional points that applying the action can add. Actions that are already complete report no additional gain.

MFA account counts are reported correctly

When identifying accounts that do not have MFA registered, Ask AI now reports the total number of accounts from the accounts it read instead of incorrectly indicating that the total is unavailable.

Most recent Defender alert dates are identified correctly

When reporting the most recent Defender alert across organizations, Ask AI now compares alert dates before identifying an alert as the latest.

Multi-organization summaries no longer conflict with detailed results

For results covering multiple organizations, Ask AI no longer opens with a summary count that can conflict with the detailed results.

Instead, it identifies what was checked and presents the details in the results. When a specific count is requested, Ask AI continues to provide the count.

Large multi-organization checks handle data limits more reliably

Checks that require large amounts of data across multiple organizations no longer automatically end with a general error when a read is too large.

Ask AI now requests only the data needed to continue the check. If the entire check cannot fit in one answer, it reports the results it retrieved and identifies the organizations or checks that were not covered so they can be addressed in a follow-up.

Mailbox checks cover more mailboxes and report inherited settings correctly

Checks of settings across mailboxes now cover several times as many mailboxes per answer, reducing cases where multi-organization checks stop before all requested mailboxes are examined.

When a mailbox follows the organization-wide setting, Ask AI now reports that behavior instead of reporting the setting as unavailable.

Billable-user exports no longer require clarification of billable status

When asked for a list of billable users, Ask AI now treats billable users as enabled accounts with at least one paid license and states that definition in its response.

Ask AI then creates the Microsoft 365 user list without requiring additional clarification. If the request refers to the billing flag maintained by SaaS Alerts, Ask AI explains that the flag is managed in the web portal while still providing the Microsoft 365-based list.