Every update to ChatGPT, Gemini, Perplexity, Copilot and Google Search can change how your brand appears in AI-generated answers. We track release notes, API changes and documentation updates across major LLM platforms so you don't have to.
Enhance agent builder with full screen mode
[Windows, Mac, Web]
Enjoy an improved agent builder with a full-screen view that streamlines the process of creating and managing your agents.
Learn more.
Enhanced adaptive card capabilities in declarative agents
Developers now have access to enhanced adaptive card functionalities—including task modules, stage view, inline actions, and charts—directly in declarative agents for richer app experiences.
Learn more.
Enhanced Q&A accuracy for SharePoint files
[Windows, Web]
Improve the precision of Q&A interactions on SharePoint files that include tables, comments, and formatting. Benefit from more relevant and insightful responses whether you're working with Word, PDF, or PowerPoint files.
Get relevant calendar results by time period
[Windows, Web]
Quickly summarize meetings for a specific day to stay on top of your schedule and focus on the events that matter most.
Improve email responses with extra context
[Windows, Web]
Get fuller email replies that expand your initial lists, indicate the number of related messages, and let you easily paginate for more details—all to help you manage your inbox more effectively.
Learn more.
Schedule meetings with smart time insights
[Windows, Web]
Easily discover optimal meeting times and streamline Outlook handoffs with intelligent calendar suggestions that make scheduling a breeze.
Learn more.
Semantically index more document pages
With improved indexing, P99 documents can now include up to 180,000 characters (about 90-100 pages) for more comprehensive document insights, while P99.99 documents support up to 1.8 million characters (about 900–1,000 pages), which is a significant increase from the previous 18-20 page limit.
Learn more.
Ground your agents with live enterprise data
[Web]
You can now enhance your copilots built with Microsoft Copilot Studio by incorporating structured data from both Microsoft and select non-Microsoft systems. These copilots enable users to ask natural language questions about enterprise systems within their Power Platform tenants. Building on the natural language query capabilities introduced with Microsoft Dataverse knowledge, Microsoft is extending this functionality to include certain third-party services.
Learn more.
Only use grounded knowledge for agent response
[Windows, Web]
Prevent agents from using model-trained knowledge by turning off internal knowledge, ensuring responses are based on specified grounded sources.
Learn more.
Onboard SharePoint Agents as a pay-as-you-go scenario in CCS
[Web]
This feature introduces SharePoint Agents to the Pay-as-you-go tab under Copilot â Billing & usage, aligning with the existing workflow used for Microsoft 365 Copilot Chat. Administrators gain the ability to manage and monitor SharePoint Agent consumption through the familiar Pay-as-you-go interface, ensuring consistent oversight across Copilot experiences. Integration with the SharePoint backend via API enables precise usage tracking and billing for this new scenario.
Learn more.
Ask Microsoft 365 Copilot for insights using Click to Do
[Windows]
Streamline your workflow by seamlessly sharing highlighted content with Microsoft 365 Copilot using Click to Do for quick insights and assistance.
Learn more.
Start an instant chat with Copilot
[Windows]
Launch a chat with Copilot quickly using the quick view for instant assistance. Use the Win+C shortcut or the Copilot key on supported devices.
Share agents with your enterprise
[Windows, Web]
Generate sharing links for your agents in Business Chat. If the recipient doesn't have the agent, they are directed to the Microsoft 365 application catalog to install it. If they do, the agent opens directly in Microsoft 365 Copilot Business Chat.
Learn more.
Analyze error codes efficiently
Use Copilot to analyze Intune error codes from device configuration profiles, compliance policies, and app installations, simplifying troubleshooting.
Learn more.
Compare device settings to uncover issues
Easily compare settings on two devices using Copilot in Intune to identify potential misconfigurations.
Learn more.
Get device summaries with Copilot
Access device-specific information such as installed apps and group membership using Copilot in Intune for better device management.
Learn more.
Copilot uses enterprise assets hosted on SharePoint OAL when creating presentations now
[Mac, Windows, Web]
Once you integrate your organization's assets into a Sharepoint OAL (Organization Asset Library) you will be able to create presentations with your organization's image.
Learn more.
Copilot uses Enterprise assets hosted on Templafy when creating presentations now
[Mac, Windows, Web]
Once you connect your asset library hosted with Templafy to Microsoft365 and Copilot, you will be able to create presentations with your organization's images.
Learn more.
Manage site ownership effectively
[Web]
The site ownership policy enables you to define and enforce ownership criteria for your SharePoint sites, automating actions to prevent data risks if sites remain ownerless for over three months.
Learn more.
Ability to Stop Copilot while it is generating a response
[Windows]
Copilot in Teams now has a 'stop' button after sending a prompt. This allows the user to stop Copilot's response either before the response has started to generate, or even after the response is generating. The user can then start a new prompt if they wish.
Interpreter agent for seamless communication
[Windows, Mac]
Interpreter Agent acts like an instant translator during your Teams meetings. It listens to the spoken language in a meeting and immediately translates it into another language in real-time. This allows participants who speak different languages to understand each other and collaborate more effectively without waiting. Whether you're holding a business meeting, customer calls, or project discussions, the AI interpreter in Teams ensures everyone can participate fully, enhancing communication and productivity across diverse teams. It supports 9 different languages: English, Italian, German, French, Portuguese (Brazil), Japanese, Spanish, Chinese (Mandarin), and Korean.
Learn more.
Enabled project permissions and sharing for Team and Enterprise plans. SeeOrganization-wide sharingfor more information.
Note:Refer toour docsfor historical release notes.
Related ArticlesHow up-to-date is Claude's training data?How large is the context window on paid Claude plans?Claude Code model configurationSet a default model for your organizationEnable US-only inference for your organization
Enabled project permissions and sharing for Team and Enterprise plans. SeeOrganization-wide sharingfor more information.
Note:Refer toour docsfor historical release notes.
Related ArticlesHow up-to-date is Claude's training data?How large is the context window on paid Claude plans?Claude Code model configurationUse Claude in Microsoft FoundrySet a default model for your organization
Enabled project permissions and sharing for Team and Enterprise plans. SeeOrganization-wide sharingfor more information.
Note:Refer toour docsfor historical release notes.
Related ArticlesHow up-to-date is Claude's training data?How large is the context window on paid Claude plans?Claude Code model configurationUse Claude in Microsoft FoundryModels, usage, and limits in Claude Code
Enabled project permissions and sharing for Team and Enterprise plans. SeeOrganization-wide sharingfor more information.
Note:Refer toour docsfor historical release notes.
Related ArticlesHow up-to-date is Claude's training data?How large is the context window on paid Claude plans?Claude Code model configurationUse Claude in Microsoft FoundryEnable US-only inference for your organization
Solve your most complex problems with Gemini 2.5 Deep ThinkWhat:Starting today, Google AI Ultra subscribers get early, limited access to Gemini 2.5 Deep Think, Gemini’s most advanced reasoning mode. Deep Think is capable of thinking for longer and generating multiple parallel streams of thought simultaneously—much like how humans brainstorm to tackle complex problems – making it excel at iterative development and design, scientific and mathematical research, and coding.Try it next time you need more brain power for your most complex tasks, especially those requiring creativity, strategic planning and step-by-step improvements. Select 2.5 Pro from the model drop down, and "Deep Think" will appear in the prompt bar. Submit your task with Deep Think enabled and Gemini will let you know when your response is ready – generally in a few minutes.Why:We're committed to bringing Google's latest AI innovations faster to Google AI Ultra users, including our most powerful reasoning capabilities to help you tackle even the most complex problems.
Enabled project permissions and sharing for Team and Enterprise plans. SeeOrganization-wide sharingfor more information.
Note:Refer toour docsfor historical release notes.
Related ArticlesHow up-to-date is Claude's training data?How large is the context window on paid Claude plans?Claude Code model configurationUsing Claude in Microsoft FoundryModels, usage, and limits in Claude Code
Enabled project permissions and sharing for Team and Enterprise plans. SeeOrganization-wide sharingfor more information.
Note:Refer toour docsfor historical release notes.
Related ArticlesHow up-to-date is Claude's training data?How large is the context window on paid Claude plans?Claude Code model configurationUse Claude in Microsoft FoundryClaude Enterprise consumption guide
- Launched image-to-video generation for the Veo 3 Preview model.
- Released Veo 3 Fast Preview model.
- To learn more about Veo 3, visit the [Veo](https://ai.google.dev/gemini-api/docs/video) page.
Study mode is currently available to users in Free, Plus, Pro, and Teams plans globally and will expand to Edu plans in the coming weeks. Study mode works with any model available in ChatGPT on iOS, Android, web, and desktop.
Study mode is a new learning experience in ChatGPT designed to help you build a deeper understanding of any topic. When you turn it on, ChatGPT will ask interactive questions to understand your goals and skill level, then work with you to reach the answer together.
With Study mode enabled, ChatGPT can:
Guide understanding with Socratic-style questions.Break concepts into easy to follow sections starting simple and adding complexity as you progress.Personalize responses based on your past chats if memory is on, using examples and tips tailored to you.Check your understanding with open-ended prompts and feedback.Work with your materials by referencing images or PDFs you upload.
You can enable study mode at any time by selectingToolsin the prompt window and choosingStudy and learnfrom the drop‑down menu or go tochatgpt.com/studymode.
Study mode is powered by custom system instructions and can have some inconsistent behavior and mistakes across conversations. We plan on training this behavior directly into our main models once we’ve learned what works best through iteration and user feedback.
Learn more about study mode.
Pro users can now connect to Canva and Notion for both chat search and deep research.
Note this feature is currently limited to users located outside of the EEA, Switzerland, and the UK.
Learn more aboutconnectors.
Pro users can now utilize chat search with HubSpot and custom connectors (MCP) in addition to deep research.
Note this feature is currently limited to users located outside of the EEA, Switzerland, and the UK.
Learn more aboutconnectors, as well ashow to create a custom connector using MCP.
Hebrew support in Agent builder
[Windows, Web]
Integrate Hebrew language support in agent builder to build accessible, localized solutions that simplify multilingual deployments and enhance user engagement.
Learn more.
Increased support for uploading up to 20 documents to agents' knowledge
[Windows, Web]
End users and makers can now upload up to 20 documents to ground agents with richer, embedded knowledge in Microsoft Copilot agent builder.
Learn more.
Share agents from embedded builder
[Windows, Web]
Users can share agents from an embedded agent builder to other individual users or group chats.
Learn more.
Share agents with context-aware link previews
[Windows, Web]
Streamline your interactions by using context-aware buttons that adapt based on where links are shared—making it easier to take the right action in chats and meetings.
Upload and embed knowledge in declarative agents
[Windows, Web]
Empower your agents with enriched context by uploading your own files and embedding crucial knowledge for personalized, day-to-day assistance.
Learn more.
Users can share agents from SharePoint
[Web]
Users can share agents via links from SharePoint to Teams group chats. This means users will be able to chat and collaborate with SharePoint agents in group chats. The shared links will unfurl into preview cards with actionable buttons to 'Add to this chat'.
Learn more.
Easily find and use knowledge data sources
[Web]
In Copilot Studio agent builder you can now quickly identify and select the right knowledge data sources without manually scanning long lists. This streamlined workflow helps you get to the insights you need faster for everyday tasks.
Learn more.
Ask Copilot to generate formulas
[Web]
Type "=" anywhere on your grid or in the formula bar and let Copilot generate formulas from natural language, making complex calculations simpler and faster.
Learn more.
Copilot advanced text analysis in Excel
[Web]
Copilot can now analyze text by identifying themes and sentiments, citing data examples, and inserting a column with labels—helping you quickly uncover actionable insights.
Learn more.
Copilot icon on the grid in Excel for the web for M365 personal and family
[Web]
Access AI-powered support directly from your spreadsheet with a single click. The Copilot icon helps you stay in your flow by offering instant insights while you work.
Learn more.
Get an audio overview of a notebook
[Web]
Turn the files in your notebook into a dynamic audio overview for an engaging listening experience. Simply select "Get audio overview" at the top of your notebook—available in English only, with more languages coming soon.
Learn more.
Dictate your prompts in Copilot Chat
[Windows, Web]
You can now use the dictation button to input your prompts via speech, making interactions with Copilot more natural and efficient.
Learn more.
Data Security Posture Management for AI - Data Risk Assessments.
[Web]
Admins can drive better security outcomes by reviewing default assessments, examining data sensitivity, and monitoring user accesses—all to quickly identify risks and remediate them in daily operations.
Learn more.
Create a presentation with Microsoft 365 Copilot from menus
[Windows]
You can quickly begin a new presentation using Microsoft 365 Copilot from the PowerPoint start menu or file menu. Copilot is easily accessible when you open PowerPoint or click the File tab, helping to simplify your workflow.
Learn more.
Microsoft 365 Copilot generates the new presentation in a new file when starting from an existing presentation
[Mac, Windows]
Now, when creating a presentation using Microsoft 365 Copilot from an existing presentation using, it creates the new presentation in a new file without affecting the original presentation.
Learn more.
Reference multiple files in your presentation creation using Microsoft 365 Copilot
[Windows, Mac, Web]
Enhance your PowerPoint presentations by referencing up to five files with Microsoft 365 Copilot, making it easier to incorporate detailed insights and comprehensive data without switching contexts.
Learn more.
Catch up on meetings with AI recaps on mobile
On iOS and Android, view AI-generated notes, tasks,
@mentions
, and speaker-indexed recordings so you can quickly get up to speed when you miss a meeting.
Learn more.
The LLM Changelog is a live, automated tracker that monitors release notes and documentation updates across the largest AI platforms — including OpenAI's ChatGPT, Google's Gemini, Perplexity AI, Microsoft Copilot and Google Search. Changes are crawled daily and presented in a single, searchable feed.
Why AI Platform Changes Matter for Brands
When AI models update their training data, adjust ranking signals or change how they cite sources, the brands and products they recommend can shift overnight. An API change to Gemini or a new feature in ChatGPT can alter which businesses get mentioned in millions of AI-generated responses.
Platforms We Track
We currently monitor ChatGPT release notes, OpenAI developer and model changelogs, Gemini API and app updates, Perplexity product and API changes, Microsoft 365 Copilot release notes, and Google Search Central documentation updates. New sources are added as the AI landscape evolves.
Built by reconnAI
reconnAI tracks how AI systems represent brands in their responses. We monitor mentions, sentiment and citations across ChatGPT, Gemini, Perplexity and more — helping businesses understand and improve their visibility in the age of AI search. Learn more →