Hello AI Enthusiast,
This week, Anthropic launched Managed Agents, a production-ready way to build and coordinate multiple AI agents without the usual infrastructure headaches. McKinsey published a 12-point AI transformation playbook that cuts through the noise on what actually works. And OpenAI's new Chief Revenue Officer shared how the company is betting its future on enterprise customers. Let's get into it.
The Big Picture 🔊
Claude Now Lets You Deploy and Coordinate AI Agents at Scale
Anthropic has launched Managed Agents, a new feature on the Claude platform that handles the infrastructure needed to run production-grade AI agents. It includes secure sandboxing, long-running sessions that keep working even through disconnections, multi-agent coordination where one agent can spin up and direct others, and built-in permission management. Companies can go from prototype to deployed agent in days rather than months. It's available in public beta on the Claude platform, priced at standard token rates plus $0.08 per active session hour.
|
||
|
This is the moment AI agents stop being demos and start becoming workplace tools. What's interesting is where this is heading inside organizations. IT departments will likely govern and manage these agent setups, setting permissions and limits. The pricing model is worth watching. Consumption-based billing means costs can grow quickly at scale, and for large companies, that predictability question matters a lot when choosing a platform they plan to rely on every day. |
McKinsey's 12-Point AI Playbook: What It Takes to Actually Transform
McKinsey has published an AI Transformation Manifesto outlining what separates companies that see real results from those that don't. The core arguments: technology alone creates no advantage, only the capabilities built around it do. Companies win when they focus AI efforts on their most strategically important processes, when senior leaders own the agenda rather than delegate it to IT, and when they treat learning as an ongoing requirement rather than a one-time event. The report also flags that adoption fails most often not because the AI is bad, but because the surrounding processes were never updated to match it.
|
||
|
What lasts is the ability to adapt, to evaluate new tools quickly, and to keep working well when the workflow changes again. That is what McKinsey calls capability building, and it is much harder to develop than a one-day workshop. |
OpenAI Is Now Seriously Chasing Enterprise Revenue
Denise Dresser, OpenAI's Chief Revenue Officer, shared a look at her first 90 days in the role. Enterprise accounts now make up more than 40% of OpenAI's revenue and are expected to reach parity with consumer by 2026. Codex, their coding agent, has 3 million weekly active users and has grown more than five times since January. OpenAI's APIs are processing 15 billion tokens per minute. The company is also positioning its Frontier platform as the enterprise AI layer, built in partnership with AWS, McKinsey, BCG, Accenture, and Capgemini.
|
||
|
A year ago, most of what we heard from OpenAI was consumer-facing: companions, creative tools, viral image generators. They are now in a direct race with Claude for enterprise customers. When enterprise revenue becomes this central to a platform's business model, the security, reliability, and governance questions tend to get answered faster. There is simply too much money on the table to leave them open. |
Bits and Bobs 🗞️
Google's Gemini AI now includes notebooks that organize chats and files to improve personalized and context-aware responses.
Gemini now offers interactive visualizations within chat, allowing users to explore complex topics through adjustable simulations instead of static images.
The 2026 AI Index highlights rapid advancements in AI capabilities, growing global adoption, shifting leadership between U.S. and China, and challenges in responsible AI development and regulation.
Claude Code now offers automated "routines" that run scheduled or event-triggered AI coding tasks with repo access and API integration.
Chrome's new Skills feature in Gemini lets users save and reuse AI prompts for faster, personalized tasks directly within the browser.
Meta's new Muse Spark model offers advanced multimodal reasoning and multi-agent collaboration, marking a key step toward personal superintelligence with efficient scaling and strong safety measures.
From Our Founder’s Channels 🤳
In his recent TikTok, our founder Gianluca Mauro shares his thoughts on whether Anthropic’s latest AI model is “too powerful” to release. Watch it here.
@gianluca.mauro Does Anthropic have an AI that is “too powerful” to release? #Ai #learnontiktok #business #claude
Most companies know what good AI adoption looks like. The hard part is actually getting there. Our Corporate AI Training helps your team build the skills and habits that last beyond the next model release, without the trial and error.
That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:
AI Academy Membership: Get 12 months of access to all our cohort-based programs and on-demand courses.
AI Agent Bootcamp: Accelerate processes and solve business problems by building AI Agents, without coding.
Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.
Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.
Catch you next week! 👋
