The phrase "AI-native" is doing a lot of work for you right now

Four things came out this week and they all said the same thing in different fonts: you're not going to get org-level transformation using the same playbooks with upgraded tools.

Lots of energy for the last year centered around redefining what's possible at the individual level (vibe coding, 10x engineers, personal AI assistants). The orgs that are furthest ahead have already moved on to redefining what's possible at the org level—and, crucially, restructuring (workflows, processes, and yes even people) around that.

Here's what we've been reading.

Every Company Can Now Be a Mini-AGI

Jack Dorsey on the Sequoia Capital podcast, as summarized by Ben Lang. Dorsey argues that traditional org charts are just "information plumbing" built for a pre-AI era. This is the company-wide version of "your existing operating model isn't cutting it anymore", so naturally we agree.

Who knows if Jack has the solution right (though Brian Armstrong from Coinbase echoed similar sentiments this week), but two things seem undeniably true: AI is already transforming information flow within the most AI-forward companies, and lots of traditional org design is based on how to best control that information flow. Do with that what you will.

OpenAI Launches the Deployment Company

OpenAI announced the OpenAI Deployment Company, a $4B+ professional services arm: Forward Deployed Engineers who will embed inside enterprises to "rethink critical operations, processes, and workflows from the ground up". Anthropic announced something similar, though with far fewer details (OpenAI acquired Tomoro to get 150 FDEs ready to deploy).

If you're still treating AI transformation as a tools rollout, you're operating at a different altitude than your competitors who are committing internal resources or bringing in FDEs.

The Truth About Being AI-Native

Greg Isenberg on what it means to actually be AI-native. If you're following along there's a theme to all these articles, and so unsurprisingly Isenberg's definition is not "individuals using AI tools". He calls that "AI-decorated."

He argues that true AI-native means the entire business is restructured so agents can operate within it, and Isenberg estimates only 1,000 companies worldwide doing $5M+ in ARR truly meet this definition.

This line gets at the core tension here: "Everyone wants the magic. Nobody wants to clean the kitchen." You have to decide whether it's rolling-up-your-sleeves time.

Learning on the Shop Floor

Tobi Lütke from Shopify. The post is about Shopify's internal AI agent River. Or maybe it's about learning in the age of AI. Or maybe it's actually about how information flows in the age of AI and the importance of rethinking your existing ideas of how companies operate to take advantage of those changes, which would certainly be convenient for the narrative we're spinning here.

The key choice Tobi highlights with River: working in the open (not default open, but literally you can only use River in public channels). This forces the kind of widespread information flow that Dorsey and Isenberg highlight as the backbone of gains from AI. Here Shopify's just using it to supercharge learning and level up their whole organization's AI fluency.

One thing to try this week

Org design is a big, messy problem that you're not going to solve overnight. But the same principles of information flow hold all the way down the org chart. What can you implement this week to start the wheels rolling on reimagining how one function, one team, or even just one workflow operates?

"Forward deploy" one of your own engineers to help unblock another function. Find a new way to work in public where both humans and AI can "see" it. Identify one workflow that an agent should be able to execute end-to-end and try it out to see where things fail.

Good luck out there. And remember, it may feel scary but it's also really, really fun.

— Claire + Zach