Excellent article John. As usual. Very enlightening.
Here is my take for entrepreneurs and operators. The opportunity is not to build a smarter AI. It’s to build the systems that let organizations trust AI with real decisions.
Here is where the real leverage is:
G — Governance of Decisions
When AI acts, when humans intervene, and how escalation works.
U — User & Identity Verification
Trust, provenance, permissions, compliance, and auditability.
A — Actionable Workflows
End-to-end ownership of real business processes.
R — Results Alignment
Clear linkage between AI use and revenue, cost, or risk.
D — Domain Context Data
Clean, rights-cleared, operationally relevant data that gives AI meaning.
Yes and i really don’t think 2026 is going to be much different. So much hype and billions of dollars being tossed around into Ai technology that will likely be obsolete next year.
The power “problem “ is different but equally fun to watch.
Fun to watch it all unfold.
Do you think that by the time we/they figure out the power issue, that there will be a new technology that doesn’t need it? And that will also become obsolete? I do 😎
The "sorting" you describe is the sound of the centrifuge spinning up.
We aren't just seeing a cultural realignment, John; we are witnessing the thermodynamic separation of the Signal from the Noise. The "middle ground" didn't dissolve because of bad rhetoric, it was rendered structurally unstable.
As I noted in The Splintering, the frictionless globalization of the last twenty years is dead. We are retreating into two distinct realities. The choice is no longer Left or Right, that is a relic of the 20th century.
The new axis is Vertical versus Horizontal. The Vertical State wants to pin you down as a dependent; the Horizontal Network lets you float.
People aren't just self-sorting. They are instinctively moving toward the only two viable positions left: total compliance or total sovereignty.
The diagnosis is accurate. Now we have to survive the treatment.
Great roundup! I’m curious how productivity gains will be distributed across industries. When consumer-facing companies like Klarna go all-in on AI and completely reengineer workflows, they often face backlash from customers and end up rolling back some or all of their AI-powered automation.
Absolutely cracking piece, John. You’ve sliced straight through the AI noise with something most execs still miss: the bottleneck isn't the tech - it's us. Loved how you framed 2025 not as a failure of models, but of management. Spot on.
Quick question though - in orgs that did shift from pilot to production last year, what’s been the most common “early win” process they redesigned successfully? Always curious what cracked first.
I think 2026 will be a massive wake-up call for people who thought tool fluency was the finish line.
The real premium is moving toward the architects who can actually redesign the workflow, not just the people who know how to write a good prompt.
Only 5% of the companies reached measurable impact… crazy low but I think will change in the coming years
This is spot on what is needed for firms to be successful with AI! There are so many golden nuggets, but the key question that you posed
“Which organizations can treat AI as a redesign exercise rather than a tool deployment?”
is the first step that leaders need to answer to be successful.
Excellent article John. As usual. Very enlightening.
Here is my take for entrepreneurs and operators. The opportunity is not to build a smarter AI. It’s to build the systems that let organizations trust AI with real decisions.
Here is where the real leverage is:
G — Governance of Decisions
When AI acts, when humans intervene, and how escalation works.
U — User & Identity Verification
Trust, provenance, permissions, compliance, and auditability.
A — Actionable Workflows
End-to-end ownership of real business processes.
R — Results Alignment
Clear linkage between AI use and revenue, cost, or risk.
D — Domain Context Data
Clean, rights-cleared, operationally relevant data that gives AI meaning.
Yes and i really don’t think 2026 is going to be much different. So much hype and billions of dollars being tossed around into Ai technology that will likely be obsolete next year.
The power “problem “ is different but equally fun to watch.
Fun to watch it all unfold.
Do you think that by the time we/they figure out the power issue, that there will be a new technology that doesn’t need it? And that will also become obsolete? I do 😎
This reads like someone who stopped measuring the year by output and started noticing what remained when the noise settled.
There’s a calm honesty here that feels earned.
A sorting year might touch each voice differently:
Nurturers notice which relationships bring life or drain it.
Guardians see which structures still serve and which don’t.
Creatives sense which ideas want to stay and which are complete.
Connectors feel circles reshaping.
Pioneers feel direction clarifying, sometimes through loss before gain.
Naming this helps people trust the process instead of resisting it.
The productivity and revenue differentials between AI leaders and laggards are already 30x on revenue per employee. Those will compound.
If this number doesn't tell you companies need to get this right, I don't know what will.
Thanks for a great piece.
The "sorting" you describe is the sound of the centrifuge spinning up.
We aren't just seeing a cultural realignment, John; we are witnessing the thermodynamic separation of the Signal from the Noise. The "middle ground" didn't dissolve because of bad rhetoric, it was rendered structurally unstable.
As I noted in The Splintering, the frictionless globalization of the last twenty years is dead. We are retreating into two distinct realities. The choice is no longer Left or Right, that is a relic of the 20th century.
The new axis is Vertical versus Horizontal. The Vertical State wants to pin you down as a dependent; the Horizontal Network lets you float.
People aren't just self-sorting. They are instinctively moving toward the only two viable positions left: total compliance or total sovereignty.
The diagnosis is accurate. Now we have to survive the treatment.
Great roundup! I’m curious how productivity gains will be distributed across industries. When consumer-facing companies like Klarna go all-in on AI and completely reengineer workflows, they often face backlash from customers and end up rolling back some or all of their AI-powered automation.
Experts who can translate individual AI implementation and gains to the organization will become very, very valuable.
Absolutely cracking piece, John. You’ve sliced straight through the AI noise with something most execs still miss: the bottleneck isn't the tech - it's us. Loved how you framed 2025 not as a failure of models, but of management. Spot on.
Quick question though - in orgs that did shift from pilot to production last year, what’s been the most common “early win” process they redesigned successfully? Always curious what cracked first.