Hey Reader,

Welcome to the 29th edition of the 3-2-1 (check out previous issues here).

One idea from me first, three links that go deeper on it, then two things that have nothing to do with work. Let's get into it.

---

1 Idea from Me

You can't steer a living company from the top

Almost every company's success is tied to how well they embrace AI in their products and the way they work.

Since things are moving so quickly, speed is critical. That means the speed of decision-making, and specifically the ability to make decisions closest to the work, without the usual routing of information to managers many levels removed.

Most leaders answer that need for speed by standing up an AI transformation office or reorganizing from the top.

Remove the limits to growth

Peter Senge's rule for the limits-to-growth pattern, from The Fifth Discipline:

"Don't push growth, remove the factors limiting growth."

Companies are like living organisms. If you steer them top down in one direction, it may not be the optimal direction for long-term survival. In 2011, Nokia's CEO Stephen Elop bet the whole phone business on Microsoft's Windows Phone, dropping its own software and passing on Android. Nokia had about 40 percent of the global handset market in 2007. By 2013 it had about 15 percent, 3 percent in smartphones, and it sold the handset business to Microsoft for $7.2 billion. Nokia chose one direction at the top and lost the market.

Instead, companies should enable adaptive capacity, the ability to run thousands of tiny experiments, to test hypotheses, gather feedback, and adjust.

Doing this well requires smart people to collaborate on these experiments, align multiple others on what feedback means, and make fast collective decisions to move work forward.

So what limits this?

Here are three big ones I've seen in my work with clients. The scenes are composites.

1. People tend to communicate how they think, not what others need to hear

An engineer spots a reliability bug in checkout and posts a long Slack update that walks through every log she checked, in the order she checked them. Her product manager skims the top, misses the pricing risk buried at the bottom, and moves on. Two days later, finance flags the same issue in a review.

When hundreds of people send one message for everyone, written in discovery order, every update needs a follow-up meeting to work out what is being asked. Decision speed drops.

Try this: map the room. Before any message that asks for a decision, have your people write down who reads it and what each reader needs, then write the version each reader needs.

2. People struggle to connect their work to the bigger picture

A designer spends two weeks refining a self-serve onboarding checklist to lift completion rates. He brings the proposal to review without talking to sales or customer success. In the meeting, an account executive points out that enterprise buyers skip self-serve onboarding entirely because their admins configure accounts in bulk.

Across a hundred experiments like that one, teams spend months optimizing local metrics that clash across functions. Work stalls in late reviews because another team spots a problem a ten-minute chat would have caught.

Try this: go where your view is weakest. Before anyone concludes, have them collect the perspective they are missing: the other function, or the person who already knows (asking for help is part of the job). One rule: your recommendation must cite one perspective you expected to disagree with.

3. People can't synthesize large amounts of information into clear opinions

A marketing manager pulls survey responses, interview notes, and campaign metrics into a twelve-page deck for a pricing test. Every slide is accurate, and none states what the team should do on Monday. After forty minutes debating slide four, the director tables the decision until next week.

Repeat that deck across hundreds of people and raw data piles up without a call. Decisions stall because the manager has to read all the raw material herself just to find the point.

Try this: sort and drop. Have your team put their raw material into three to five non-overlapping buckets, then drop what will not change the decision. Dropping is the skill. Make them name what they dropped and why, then state the recommendation in the first sentence.

Your move

Pick the limitation your team hits most and give the team one question to ask each other before work reaches you:

• For a message written in discovery order: "Who is this written for?"
• For work done in a silo: "Who outside your team did you ask?"
• For a data dump: "What did you drop?"

Hit reply and tell me which of the three your team hits most.

---

3 Things for Work (in L&D)

How frontline empowerment and AI reshape organizational performance and enable growth (Roland Berger, Sep 2026)

• Their research: 46% of companies still run three or more layers between the board and the teams facing the market: "decision authority has not moved with the information." Four levers follow, including explicit decision rights.
• 🏋 Effort ≈ 8 min read

Decision Latency: Why AI Productivity Gains Stall Between Headquarters and Global Teams (Zinnov, Sep 2026)

• Make the build three times faster and the whole cycle speeds up only 2.1x, because the waiting didn't change. The number to watch is decision latency, the time work sits idle waiting for an approval. Written about offshore teams; it applies to any team.
• 🏋 Effort ≈ 5 min read

Distributed Intelligence: Why Hierarchies Can't Keep Up (Jonathan Brill, Jul 2026)

• The optimistic case, with examples: Uber embedding AI engineers in finance and operations, a Bayer line run by five people instead of fifteen. His measure is decision velocity, how fast you spot a problem, decide, and learn from the result.
• 🏋 Effort ≈ 8 min read

---

2 Things for Life

Tributary (tributary.music)

• A fun rabbit hole of musical influence. Start with an artist you like and trace the musicians who shaped them, with the songs and the stories behind each link. I've been getting into modern country and roots music, and it's been fun to track the earlier influences on my favorite artists.
• 🏋 Effort ≈ as long as you let it run

Kids Trends (Riley Walz)

• Twenty years of childhood, one drawing at a time. My 6-year-old is a budding artist, and this was a fun way to track the popular things kids draw. Rocket ships shot up in the early 2010s before tailing off dramatically, while jellyfish are weirdly growing in popularity.
• 🏋 Effort ≈ 5 min of scrolling

---

That's it for this week - enjoy your Sunday!

I'll be back in two weeks ✌️

Andrew

P.S. 👉 If you lead a talent or L&D team and the topic of judgment is coming up a lot, hit reply and let me know. I'm working on something to bring perspectives together.