Hope your week is going well! We’ve got a good one this week. There's been a lot happening in the AI space but here’s my curated list of updates worth knowing about.

I'm also diving into new Claude capabilities that I think could change the way your team uses Claude day to day.

This week we're covering:

OK, let's get into it!

Claude just rolled out another great new feature. Anthropic has been building out Claude in ways that really matter for how we work day-to-day. Not just making the model smarter, but adding the infrastructure around it that turns it into a very helpful business tool.

First, the base model is solid. It handles a very large context window (hundreds of pages), which means it can work with more information at once without losing track.

But the real value is in what they've built around it:

Projects let you create dedicated workspaces with custom instructions and a “knowledge base” for context that persists across conversations. Set up a project for each client, every big project, or for every area of your business. Claude regenerates project memory every evening from your past chats in that project.

Voice Mode on mobile lets you have full spoken conversations with Claude. Brainstorm while walking, dictate ideas while cooking, or ask questions when your hands are busy.

Claude Code to  help you build automations and scripts even if you're not a developer.

And now: Cowork Plugins

This is the new announcement from this week, and it's the piece that brings everything together.

This matters most if you're supporting sales, ops, marketing, or leadership teams that repeat the same workflows every week.

Cowork plugins let you bundle skills, connectors, slash commands, and sub-agents together so Claude becomes a specialist for your specific role and team. Not just a general chat tool. An actual specialist built around how you work.

Plugins have three main parts that work together:

Skills are like Claude's built-in instincts for specific tasks. When you install a plugin, it comes loaded with expertise for that domain. For example, a sales plugin knows how to research prospects, qualify leads, and draft outreach emails. You don't have to tell Claude to use these skills. If you're working on something and Claude recognizes a relevant skill from your installed plugins, it just uses it automatically in the background. 

Commands give you direct control when you need it. These are shortcuts you trigger yourself by typing a slash and picking from a menu. Think of it like keyboard shortcuts, but for AI workflows. 

Connectors are the bridge between Claude and your existing tech stack. They pull data from the tools you already use: your CRM, email, Slack, Google Drive, etc and make it available to Claude without you having to copy and paste or switch between apps. 

Skills work in the background, commands you trigger yourself, and connectors tie everything together.

And here's the part I think is really useful. Anthropic is open-sourcing 11 plugins that their own team already uses. So you're not starting from zero. They cover sales, finance, legal, marketing, customer support, data analysis, product management, and enterprise search

And there's one more I want to call out. The Plugin Creator plugin creates other plugins. You describe your workflow in plain English, and it builds the skills and commands needed to run it. 

It's available right now as a research preview for all paid Claude users. 

If your team is already using Claude, this is a simple way to get a lot more out of the tool. 

My recommendation if you're getting started:

Pick one plugin that matches your most repetitive workflow and test it to see how it actually performs. 

Let me know if you try any of this out. I'd love to hear what works for you.

OpenAI just announced they're putting ads in ChatGPT. Anthropic saw the opening and responded with one of the smartest ad campaigns I've seen in a while.

The ads are actually hilarious. Each one shows someone asking ChatGPT a genuine question - how to get in shape, improve communication with their mom, finish an essay, or start a business. ChatGPT starts giving helpful advice, then suddenly pivots to pushing ridiculous sponsored products.

It's satire, but it makes the point crystal clear: ads could seriously disrupt the AI assistant experience.

I use both ChatGPT and Claude for different things, and I'm curious to see how ads actually get implemented. Will they be subtle and non-intrusive? Or will they break your flow when you're in the middle of solving a business problem?

The reality is we don't know yet how OpenAI will handle this. But Anthropic is making a clear bet that "no ads" is a competitive advantage worth advertising.

Here's what this signals: For teams using AI for sensitive work, client strategy, or confidential analysis, the "no ads" commitment matters. It's about trust and focus when you need it most.

Watch the full ad series here - they're worth a few minutes of your time.

I'm genuinely curious how much AI was used to create these. The production quality and comedic timing are excellent.

That's it for this week. Hopefully something in here sparked an idea or saved you a little time. If anything stood out or you want to dig deeper on any of it, you know where to find me. 

See you next week,

Julien

PS: If this was helpful, forward it to someone on your team who'd get something out of it. Always appreciate the support www.ampra.ai/join-our-newsletter.