AI · MICROSOFT · SHARED KNOWLEDGE

AI, in your work.

Understand what’s changing.
Discover what you can do with it.

IN FOCUS / AI & industry

Copilot Chat limits: preparing support teams and user expectations

Microsoft extends usage limits to additional Copilot Chat features for users without a Microsoft 365 Copilot license in commercial organizations. The notifications call for support preparation and a review of license requests, without turning every limit into a recommendation to buy.

Read the story 3 min

TRY IT IN NEXO

A useful answer starts before the question.

Choose a task. Compare two requests. Take away a prompt with context.

Open the prompt workshop Interactive example · Fictional data · No signup
FROM IDEA TO NEXT STEP

“Summarise this meeting.”

ContextGoalFormat

“Extract decisions, actions and what still needs confirming.”

The difference is in what you ask.

EXPLORE NEXO

The latest, with perspective.

73 articles

Topic: All

Agent 365: reviewing identity, ownership and governance actions

Agent 365 has been generally available commercially since May 1, 2026, with per-user licensing; the sources consulted do not specify Action Center’s rollout status.

Microsoft · Announcement: · Editorial review

Microsoft Agent 365 overview with an agent inventory, usage charts, and cards about risks and ownerless agents.
Microsoft Agent 365 dashboard (2025 archive). © Microsoft · Used with permission from Microsoft.

Copilot across environments: Xcode, JetBrains Air and PR tracking

Copilot comes to Xcode's built-in coding assistant in public preview and to JetBrains Air. Agent Merge also improves pull request tracking, with different conditions depending on the environment.

GitHub · Announcement: · Editorial review

Laptop showing Xcode with Swift code in the editor and an iPhone app preview.
Xcode 27. Image: Apple.

Sales orders into BC: checking compatibility with recurring lines

A BCApps commit describes a numbering conflict when importing Dynamics 365 Sales orders into Business Central with automatic recurring lines, relevant to acceptance testing for the sales integration.

Microsoft · Announcement: · Editorial review

A Microsoft Dynamics 365 Business Central order on the left and a Microsoft Dynamics 365 Sales order on the right, with their product lines.
Orders in Microsoft Dynamics 365 Business Central and Microsoft Dynamics 365 Sales. © Microsoft. Used with permission from Microsoft.

CodeQL and Xcode 27: planning continuity for code scanning

Restrictions involving macOS 27 and Xcode 27 require a review of code extraction. Analysis without a build offers alternatives with limits that depend on the language.

GitHub · Announcement: · Editorial review

MacBook Pro displaying a code editor.
Programming on a MacBook Pro (archive). Photo: Arnold Francisca / Unsplash.
nNexo CommunityEditorial review · 2026-11-08

ANNOUNCEMENT & AVAILABILITY

Official documentation checked on November 8, 2026. Coverage describes compatibility by language and extraction method. The example is hypothetical; no workflows from specific repositories have been tested.

CodeQL changes for macOS 27, Xcode 27, and Swift 6.4

What changes

In its October 19, 2026 notice, GitHub links the restrictions to the removal of Intel code from binaries and libraries needed for build tracing.

GitHub recommends pinning macOS 26 or earlier and Xcode 26 or earlier for Swift and other scans using build tracing.

ChangeReported effect
Xcode 27 installedNo traced extraction for C/C++, C#, Go, Java, Kotlin, or Swift.
macOS 27Traced extraction may be unavailable.
Intel-based MacSupport deprecation in mid-December 2026.

Why the extraction method matters

CodeQL represents code in a database and runs queries against it to detect vulnerabilities and coding errors. It can run through GitHub Actions or integrate with other continuous integration systems, according to the CodeQL CLI documentation.

When tracing is involved, CodeQL intercepts build processes and forwards information to each language’s extractors, as the technical build reference explains. Nexo’s interpretation: successfully building a project and successfully extracting its code for analysis are separate checks.

What analysis without a build offers

The compiled languages guide provides `build-mode: none` for C/C++, C#, and Java: it creates the database without building the entire project. Extraction can become less accurate when dependency information or code generated during a build is missing.

The current platform requirements limit macOS 27 support to Apple Silicon and build mode `none`. Changing processors alone therefore does not resolve the tracing restrictions.

  • C/C++: compilation options and header paths are inferred. Custom macros and external dependencies can reduce accuracy. C#: problems restoring dependencies can also affect analysis. Language-specific details.
  • Java and Kotlin: with `none`, Kotlin code is not analyzed and a warning is produced. A mixed repository needs to retain a build-based mode to include Kotlin. Limits of build mode none.
  • Swift: the compatibility matrix lists Swift 5.4–6.3 and requires macOS; Swift 6.4 falls outside the published range. The documented modes are `autobuild` and `manual`, with no `none` alternative. Build options.

Pinning the image does not freeze every tool

The GitHub Actions runner reference offers `macos-26` for arm64. The inventory for that image, checked on November 8, 2026, lists Xcode 26.6 as the default.

GitHub explains that `-latest` labels can migrate to other operating system versions and that images receive weekly updates. Choosing a versioned label avoids following that migration, but does not freeze all installed software. Image and update policy. As a practical measure, record the image and actual Xcode, Swift, and CodeQL versions for each run.

A hypothetical example and practical checks

Consider a repository containing a Swift application and a Java service without Kotlin. One planning option would be to retain a compatible environment for Swift analysis and evaluate `none` only for Java, reviewing its dependencies and generated code. This illustrates decisions by language; it does not report results from a tested migration.

Based on those differences, Nexo proposes reviewing each integration job with these checks:

  • Identify the language, operating system, architecture, and CodeQL version.
  • Record the installed Xcode versions and the Swift version actually used by the analysis.
  • Document the extraction mode and any dependencies or generated files the project needs.
  • Check extraction logs, uploaded results, and the scope of analyzed code before and after the change.

Created with AI assistance and reviewed by Nexo Community. Practical proposals and editorial interpretation are distinguished from documented facts.

Read the source ↗

CONNECT THIS IDEA

A useful next step.

AI & industry · 2 minAgent 365: reviewing identity, ownership and governance actionsAI & industry · 4 minBCApps adds control over order quantities after credit memos and corrections

More articles appear as you continue scrolling.