Jon W. Hansen, FCIPS | Procurement Insights | Hansen Models™
AI can copy anything that is written down. The strands that decide your outcomes were never written down.
In October, a developer and a small team used AI coding tools to build early versions of seven Adobe applications in a matter of days. They are far from finished; by the project’s own estimate, only two are genuinely usable so far. But a credible start on software that took decades to build now takes days.
The conclusion many people are drawing is that software is no longer an advantage. That is true. But it was never the advantage in the first place.
The advantage is seeing the strands.
Where it started: 1983
In 1983, my first year in technology, I was programming in dBASE II. I noticed something I could not explain at first. I would change one line of code, and another line, somewhere else, that had been working perfectly, would stop working. The two lines seemed to have nothing to do with each other. But they were connected.
That was the first time the pattern crossed my mind: things that look unrelated are often connected, and the outcome is decided by how they connect. I formulated it as a theory in 1998, applied it in the field, and named it in print in 2008: Strand Commonality™.
What a strand looks like
In 1998, a parts operation was delivering next-day only about half the time, against a 90 percent target. The software was not the problem. Tracing back from the outcome, the first strand was simple: orders were reaching the system at 4 p.m. Why 4 p.m.? Because the service department held its parts orders until the end of the day, and its staff were measured on calls answered. By 4 p.m., next-day delivery was already out of reach.
That strand led to the next: suppliers with no experience clearing goods across the border. Then the next: finance recording the cost of the part but not the cost of shipping it, so the operation was losing money on every transaction it thought was profitable.
None of those strands were in the software. None were on the process map. A developer, or an AI, could have copied the ordering system perfectly and reproduced every one of those problems.
You cannot successfully automate an operation you do not understand. You have to understand how it actually runs first, not how it is documented, and only then build or automate. That is where Strand Commonality™ comes in. It is not a matter of coding. It is a matter of understanding.
The process map can be copied in days. The strands beneath it cannot be seen until someone traces them: copy the map, and you copy every problem hiding under it. Illustrative.
Why AI cannot copy this
AI learns from what is written down: code, documents, data, the documented process. Strands live where nothing is written down: in how people are measured, when things actually happen, who has the authority to act, and the workarounds no system records. You find them by starting from the outcome and tracing back, and each one you find reveals the next.
And knowing which strands matter takes something AI cannot generate after the fact: a record of how the same strands have crossed before. In 2006, in my Summit Magazine column, “Technology’s Diminishing Role in an Emerging Process-Driven World” (referenced in my June 2007 post), I wrote that “process, not technology, is the main force behind successfully achieving efficiency and spend rationalization.” The Procurement Insights archive has followed the same strands through nineteen years of dated writing since. Our labs lock their expectations before events unfold, then score them against what actually happens, including the misses.
AI compresses the time it takes to build a capability. It does not compress the time it takes to learn how the strands behave. I first saw the pattern in 1983, and in 1998 it was proven in a live production environment, with Canadian government research funding behind it. The strands are there whether you can see them or not. Just because you cannot see them does not mean they are not shaping your outcome.
AI has changed what I can see. Where I once worked with two, three or four strands at a time, I can now see fifteen to twenty at once. But seeing more strands is not the same as recognizing which ones matter, and how they connect. That recognition is the real skill, and it is the heart of succeeding with AI.
What this means for you
If every competitor can soon build the same software, the software will not separate you from them. What will is whether you can see the strands your decisions sit on: the incentive that delays the order, the authority nobody holds, the cost nobody records.
Those strands live in the shadow operations: the workarounds, the spreadsheets that are still in use, and the tools nobody approved. In a March 2026 survey of 2,001 US and UK employees, nearly two-thirds of senior decision-makers admitted to using unapproved AI tools, more than twice the rate of the people they manage. That is a shadow operation, and it starts at the top. If you do not capture it, you are working from a clean, neat, packaged process map that is not real.
Strand Commonality™ changes how that decision is made. Most technology decisions start from the technology: the initiative is to make the tool work in your environment. Strand Commonality™ starts from the other end. It makes you look beyond the boundaries of your operation, inside your organization and outside it, to understand each strand on its own and what the strands do together. Only then is the technology chosen, to fit how your world actually works.
So before you choose a technology, don’t ask “What can it do?” Ask “How does our world really work, and will this fit it?”
This is also what AI governance really is. It is not about coding, and it is not, at heart, about guardrails or harnesses. Those matter, but they are written for the process as documented. AI governance is about understanding the strands: how the unique operating attributes of competing strands connect and affect one another. Guardrails built without that understanding guard the map, not the operation.
Next: 2000 and 2026, side by side
In my next post, I will show the same strands at work twenty-six years apart: a fire at a chip plant that hit Nokia and Ericsson in March 2000, and a freezer alarm at 3 a.m. in 2026. The technology is completely different. The strands that decided both are not: who had the authority to act, and how quickly they used it.
Start with one decision
Before you commit to an AI platform or a major software change, run the decision through Hansen’s AGR Snapshot, our two-week readiness test. We trace the strands you are not yet seeing and pressure-test the decision before the alarm sounds. At the end, you receive a decision-authority map, a pressure-test finding and a readiness reading: proceed, fix first, or stop.
The Snapshot needs no system access and only limited, agreed time from your team. The fee is fixed and credited in full to the next stage.
Truth Is Believing. Accuracy Is Knowing. Outcome Is Proof.™
-30-
Sources
Reporting on Brandon Thomas’s ArtCraft project: TweakTown, 8 October 2026; How-To Geek, October 2026
TrustedTech and Censuswide, Shadow AI in the Workplace, March 2026 (2,001 UK and US employees; 65% of senior decision-makers vs. 31% of other employees), as reported by CSO Online
Procurement Insights, Dangerous Supply Chain Myths (Part 4), 6 June 2007, referencing “Technology’s Diminishing Role in an Emerging Process-Driven World” (Summit Magazine, 2006)
AI Can Copy the Software. It Cannot See the Strands.
Posted on November 9, 2026
0
Jon W. Hansen, FCIPS | Procurement Insights | Hansen Models™
AI can copy anything that is written down. The strands that decide your outcomes were never written down.
In October, a developer and a small team used AI coding tools to build early versions of seven Adobe applications in a matter of days. They are far from finished; by the project’s own estimate, only two are genuinely usable so far. But a credible start on software that took decades to build now takes days.
The conclusion many people are drawing is that software is no longer an advantage. That is true. But it was never the advantage in the first place.
The advantage is seeing the strands.
Where it started: 1983
In 1983, my first year in technology, I was programming in dBASE II. I noticed something I could not explain at first. I would change one line of code, and another line, somewhere else, that had been working perfectly, would stop working. The two lines seemed to have nothing to do with each other. But they were connected.
That was the first time the pattern crossed my mind: things that look unrelated are often connected, and the outcome is decided by how they connect. I formulated it as a theory in 1998, applied it in the field, and named it in print in 2008: Strand Commonality™.
What a strand looks like
In 1998, a parts operation was delivering next-day only about half the time, against a 90 percent target. The software was not the problem. Tracing back from the outcome, the first strand was simple: orders were reaching the system at 4 p.m. Why 4 p.m.? Because the service department held its parts orders until the end of the day, and its staff were measured on calls answered. By 4 p.m., next-day delivery was already out of reach.
That strand led to the next: suppliers with no experience clearing goods across the border. Then the next: finance recording the cost of the part but not the cost of shipping it, so the operation was losing money on every transaction it thought was profitable.
None of those strands were in the software. None were on the process map. A developer, or an AI, could have copied the ordering system perfectly and reproduced every one of those problems.
You cannot successfully automate an operation you do not understand. You have to understand how it actually runs first, not how it is documented, and only then build or automate. That is where Strand Commonality™ comes in. It is not a matter of coding. It is a matter of understanding.
The process map can be copied in days. The strands beneath it cannot be seen until someone traces them: copy the map, and you copy every problem hiding under it. Illustrative.
Why AI cannot copy this
AI learns from what is written down: code, documents, data, the documented process. Strands live where nothing is written down: in how people are measured, when things actually happen, who has the authority to act, and the workarounds no system records. You find them by starting from the outcome and tracing back, and each one you find reveals the next.
And knowing which strands matter takes something AI cannot generate after the fact: a record of how the same strands have crossed before. In 2006, in my Summit Magazine column, “Technology’s Diminishing Role in an Emerging Process-Driven World” (referenced in my June 2007 post), I wrote that “process, not technology, is the main force behind successfully achieving efficiency and spend rationalization.” The Procurement Insights archive has followed the same strands through nineteen years of dated writing since. Our labs lock their expectations before events unfold, then score them against what actually happens, including the misses.
AI compresses the time it takes to build a capability. It does not compress the time it takes to learn how the strands behave. I first saw the pattern in 1983, and in 1998 it was proven in a live production environment, with Canadian government research funding behind it. The strands are there whether you can see them or not. Just because you cannot see them does not mean they are not shaping your outcome.
AI has changed what I can see. Where I once worked with two, three or four strands at a time, I can now see fifteen to twenty at once. But seeing more strands is not the same as recognizing which ones matter, and how they connect. That recognition is the real skill, and it is the heart of succeeding with AI.
What this means for you
If every competitor can soon build the same software, the software will not separate you from them. What will is whether you can see the strands your decisions sit on: the incentive that delays the order, the authority nobody holds, the cost nobody records.
Those strands live in the shadow operations: the workarounds, the spreadsheets that are still in use, and the tools nobody approved. In a March 2026 survey of 2,001 US and UK employees, nearly two-thirds of senior decision-makers admitted to using unapproved AI tools, more than twice the rate of the people they manage. That is a shadow operation, and it starts at the top. If you do not capture it, you are working from a clean, neat, packaged process map that is not real.
Strand Commonality™ changes how that decision is made. Most technology decisions start from the technology: the initiative is to make the tool work in your environment. Strand Commonality™ starts from the other end. It makes you look beyond the boundaries of your operation, inside your organization and outside it, to understand each strand on its own and what the strands do together. Only then is the technology chosen, to fit how your world actually works.
So before you choose a technology, don’t ask “What can it do?” Ask “How does our world really work, and will this fit it?”
This is also what AI governance really is. It is not about coding, and it is not, at heart, about guardrails or harnesses. Those matter, but they are written for the process as documented. AI governance is about understanding the strands: how the unique operating attributes of competing strands connect and affect one another. Guardrails built without that understanding guard the map, not the operation.
Next: 2000 and 2026, side by side
In my next post, I will show the same strands at work twenty-six years apart: a fire at a chip plant that hit Nokia and Ericsson in March 2000, and a freezer alarm at 3 a.m. in 2026. The technology is completely different. The strands that decided both are not: who had the authority to act, and how quickly they used it.
Start with one decision
Before you commit to an AI platform or a major software change, run the decision through Hansen’s AGR Snapshot, our two-week readiness test. We trace the strands you are not yet seeing and pressure-test the decision before the alarm sounds. At the end, you receive a decision-authority map, a pressure-test finding and a readiness reading: proceed, fix first, or stop.
The Snapshot needs no system access and only limited, agreed time from your team. The fee is fixed and credited in full to the next stage.
What strands are you missing? Contact us at HPT@hansenprocurement.com to review one decision.
Truth Is Believing. Accuracy Is Knowing. Outcome Is Proof.™
-30-
Sources
Share this:
Like this:
Related