Hello AI Enthusiast,
This week, the most useful AI news came from admin panels and update logs. Microsoft packed Copilot with new features, and Anthropic gave companies a clearer view of what their teams do with Claude, and what it costs.
Let's get into it.
The Big Picture 🔊
Microsoft Packs Copilot With Agents and Skills
Microsoft's October Copilot update brings Chat and Cowork into one Home screen and introduces Autopilot, an always-on agent with its own mailbox and files that works toward goals users set. Skills come to Copilot Chat, and PowerPoint users can record a task to save it as a skill. SharePoint gets a knowledge base built from emails, meetings and files. Most features reach early-access Frontier users first.
|
||
|
This might be the first Copilot update that shows Microsoft knows how to make money with AI. None of these features is new, but Microsoft turns them into buttons the average employee can press, inside the Word, Excel and Outlook they already use. OpenAI and Anthropic can't match that. The weak spot is price, since an always-on agent can burn through credits fast. Good features don't help if companies are afraid to switch them on. |
Claude Enterprise Shows Where the AI Budget Goes
Anthropic added three admin tools to Claude Enterprise. Smart reports analyze a sample of a team's conversations, group the work into workstreams, show what each one costs, and recommend which recurring tasks to turn into skills. In-app surveys ask employees what value Claude gives them. Pooled group budgets give each team one shared monthly budget, so heavy users don't hit their personal limit early.
|
||
|
Our favorite part: smart reports spot the tasks people repeat and recommend turning them into skills. The rest is about proving value, which also gives Anthropic a strong argument when prices go up after the IPO. Still, this is what Anthropic does well: small, steady updates that add up to a product companies actually use. |
Bits and Bobs 🗞️
OpenAI published an updated Model Spec, the document that describes how its models should behave, asking them to use judgment beyond literal instructions.
OpenAI is merging ChatGPT's Chat and Work tabs into one space, starting with select users. ChatGPT and Codex will also share one usage allowance.
Google now lets users generate, preview, and edit files like PDFs, Docs, and slide decks directly in the Gemini app.
Anthropic updated its IPO filing with third-quarter revenue of nearly $17 billion. The filing also adds a new risk: Anthropic says it may soon meet its own threshold for AI that dramatically speeds up its own development.
Google introduced Gemini 4 Carbon, a model built for cybersecurity and long engineering tasks, available first to a small group of cyber defenders.
Anthropic launched artifacts for legal work, ready-made interactive tools for tasks like contract review that legal teams can adapt and share.
OpenAI proposed a shared crawler signal to make it easier for website owners to say whether their content can be used for AI training.
Google Workspace added Markdown editing in Docs and live speech translation in Meet, plus AI meeting summaries and follow-ups.
Giving employees AI tools is the easy part. Knowing where they create value, and where they just burn credits, is harder. Our Corporate AI Training helps teams use AI where it actually pays off.
That's a wrap on our newsletter!
If you're looking to build AI skills in your team, check out our Corporate Training. Every program is tailored to your team and built around hands-on practice.
Catch you next week! 👋
