Designed for Behavior

How the principles of human change should shape every decision in building a company

My undergraduate degree is in music therapy. Four years of studying behavior, and specifically how to influence it. I did not go directly from music therapy into enterprise technology. But I have never left that foundational understanding behind. It has shaped every significant professional contribution I have made since, including the invention of Customer Success Management at Vantive in 1995, and now the company and platform we are building at Vision Intelligence.

This piece explains how. It is a companion to the thought leadership essay on cognitive dissonance and successive approximation. That piece makes the behavioral argument in the abstract. This one tells the story of where the argument came from, and how it became a company.

What Music Therapy Teaches About Change

Music therapy is not what most people imagine. It is a clinical discipline that uses music as a tool for behavioral modification. The music itself is almost incidental. What is being designed is the behavior and the experience.

Give a person with low confidence an autoharp: everyone can produce a good sound from one immediately, without training. That first success is not about music. It is about the experience of capability arriving before the person believed it was possible. Form a group struggling with social interaction into a choir: the behavior required to sing together, to listen, to time your contribution to fit the whole, is the social skill itself, learned through doing rather than instruction. Claus Bang used music therapy and resonator bells to teach deaf people to speak with inflection. The applications are wide. The underlying principle is consistent: the tool creates the conditions for a behavior to occur, and the behavior changes what the person believes is possible for them.

That is the principle I have built every significant piece of work on since. You cannot impose a new state on a person. You can only create the conditions in which they move toward it themselves. Start where people are, not where you want them to be. Make the first step achievable. Let the experience of that first success change the belief. Then build from there.

The technical term for this sequence in behavioral psychology is successive approximation toward a goal. The practical consequence is this: if you change the reality before you change the belief, the person will work to bring the reality back in line with what they believe is true. Resistance is not stubbornness. It is the mind doing exactly what it was designed to do.

From Music Therapy to Customer Success

When I joined Vantive in the mid-1990s, enterprise software was failing its customers at a remarkable rate. The prevailing assumption was that this was a training problem. What I saw was a sequencing problem. Implementations were changing the reality before they changed the belief, and customers were doing what people always do: working to close the gap by reverting to what they already knew.

But the behavioral work had to begin inside Vantive before it could reach Vantive's customers. I remember telling new hires exactly this: my goal is to create something special in terms of the level of service we provide our customers, and I know that my job is to create the something special in our corporate environment in which you can create the something special for them. The customer's experience had to be designed from the inside out.

This is the part of Customer Success Management that is most frequently misunderstood. CSM is not a customer-facing role. It is a company-wide operating model. The single point of contact with the customer will always fail if the company behind it is not designed to support it. The product team whose decisions create friction. The finance team whose invoicing creates confusion. The engineering team whose choices determine what is possible. Every team, every function, every person has a role in the customer's experience, whether or not they ever speak directly to a customer. Building CSM at Vantive meant changing how the entire company thought about its relationship with the customer, one team at a time, one visible success at a time.

The reason CSM has four million practitioners but most implementations still disappoint is that organizations adopted the role without adopting the philosophy. They changed the last mile without changing the road.

The Engineer Who Got There First

I want to tell you about my the founder of Vision Intelligence, Dr. Krste Pangovski.

Krste built the foundational architecture of the VIOLET platform before we met. He designed the modular detector library, the Standard of Work engine fully decoupled from detection, the edge-native processing model, and the open platform that allows any model to run alongside VIOLET's own. He made these decisions on purely technical grounds, driven by a deep understanding of what computer vision architecture needs to look like when it is genuinely trying to solve the adoption problem rather than just the detection problem.

When I encountered his work, my reaction was something I still find difficult to describe precisely. He had arrived at the behavioral answer from the technical direction. Every architectural decision, made without any background in psychology or human change, was precisely what a behavioral scientist would have designed if they were trying to build a platform that humans could actually adopt.

I have been known to say, with affection and not entirely in jest, that Krste is just short of brilliant, because he did not know at the time just how brilliant his solution was on the behavioral level. He solved the human problem without knowing he was solving it. That, I think, is what genuine insight looks like.

What we built together is the convergence of two lines of thinking that arrived at the same destination from opposite directions. I came from psychology, human behavior, and three decades of customer success work. Krste came from first principles of computer vision architecture. Neither of us compromised to meet the other. The architecture he designed and the philosophy I brought are expressions of the same underlying truth about how change actually works.

The VIOLET platform was not retrofitted with behavioral principles after the fact. The behavioral principles and the technical architecture are the same thing, expressed in different languages.

The Human at the Center

We are building this in a particular moment. Artificial intelligence and robotics dominate the conversation about the future of manufacturing. The narrative is largely one of replacement: machines doing what humans once did, algorithms making decisions humans once made, automation eliminating the variability that human beings introduce.

That narrative is not wrong, exactly. Robots have taken the structured, repetitive, physically demanding work. The low-hanging fruit has been picked. Automated systems handle the tasks that are most amenable to automation: those with high regularity, low variability, and clearly defined success criteria.

But what remains, and what will remain for a long time, is fundamentally human. The skilled assembler building a complex aerospace component. The experienced operator who knows by feel when a process is running slightly wrong before any sensor has registered a deviation. The trainer who transfers twenty years of craft knowledge to a new hire. The engineer who looks at a pattern in the data and asks the question that no algorithm thought to ask. These are not things automation has failed to reach yet. They are things that matter precisely because they are human.

The behavioral dimension of what we are building is not incidental to the technology. It is the point of it. VIOLET observes human work and learns from it. It captures the knowledge of the skilled operator before that knowledge walks out the door. It supports the new hire in developing capability at their own pace, with guidance that meets them where they are rather than demanding they perform at a standard they have not yet reached. It measures variation not to expose failure but to identify excellence and make it transferable.

This is, I think, a genuinely different relationship between technology and the human being at the workstation than the one most people imagine when they hear the words artificial intelligence and manufacturing. The technology is not there to replace the human. It is there to see them, to learn from them, to protect what they know, and to help the next person become what they became.

In an era when the dominant question about AI seems to be how much human work it can eliminate, Vision Intelligence is asking a different question: how much human knowledge can we preserve, transfer, and build on? That is a behavioral question before it is a technical one. And it is the reason the behavioral design of this platform is not a philosophy we added after the architecture was built. It is why the architecture exists at all.

Robots have picked the low-hanging fruit. What remains is human, and human dignity in manufacturing matters. VIOLET is built for that reality.

The Convergence

I studied music therapy because I wanted to understand how human beings change. I built Customer Success Management because I watched enterprise software fail to understand that the human change was the harder problem. I co-founded Vision Intelligence because I met an engineer who had, without knowing it, designed a technical architecture that solved the human problem from the inside out.

What I want to be clear about, as this piece closes, is that the behavioral design of Vision Intelligence operates at three distinct levels simultaneously. They are not three separate applications of the same principle. They are the same principle at three different scales.

The first is the company itself. Vision Intelligence was designed behaviorally from the inside out, in exactly the way CSM was built at Vantive. The commercial model, the deployment sequence, the way the team was assembled, the way we approach every customer relationship, all of it reflects the same understanding: that you have to create the right conditions inside before you can create the right conditions outside. The company is an expression of the philosophy, not just the product.

The second is the adoption journey. The successive approximation deployment model is designed to modify the belief system of the people adopting the technology. Each deployment is calibrated to deliver a win that shifts the belief just enough to make the next step feel natural rather than threatening. The cognitive dissonance that normally resists change is redirected. Instead of pulling people back toward the familiar, it begins pulling them toward the platform, because the platform is now consistent with what they believe is possible. Adoption is not something we ask for. It is something the deployment sequence produces.

The third is the process itself. The platform is designed to modify the behavior of the people doing the work. Not by surveillance, not by enforcement, not by imposing a standard from above, but by creating the conditions in which better behavior emerges from within. The operator who is guided through a complex assembly step does not feel monitored. They feel supported. The new hire who reaches competency faster than anyone expected does not feel measured. They feel capable. The team whose process variation is made visible does not feel exposed. They see, for the first time, what excellence in their own operation actually looks like, and they move toward it.

This is the convergence that gives Vision Intelligence its coherence as a company. The same principle, operating at the level of the organization, the adoption journey, and the individual human being at the workstation. Behavior first. At every scale.

That is what we are building. Not just a computer vision platform. An architecture of possible belief.