Organ, vitamin, or cancer: modeling a business into a system
The fuzzy ask 'make this business AI-first' modeled into an operating system: three laws (organ not vitamin, organ not cancer, offer over client) that decide what to build and what to cut.
“Make this business AI-first” means nothing until you model it. It is a mood, not a spec. The work is turning that mood into a structure that decides, on its own, what to build and what to cut.
Two ways to fail before the model
- The consultant diagnoses and hands over a beautiful slide. A map with no engine. Nothing moves.
- The implementer automates without governance and increases the speed of the error. Automation amplifies whatever it points at, so pointing it at chaos just accelerates the chaos.
The third way is to install a governed operating system: it moves fast, in the right direction, and it keeps running without you. That “governed” is the whole game, and it comes from three laws.
The three laws that do the deciding
Organ, not vitamin (fit). Does the offer become load-bearing, so the client’s revenue runs through it and turning it off hurts fast and visibly? Or is it a removable vitamin that churns by month five? The real ICP filter is not the vertical, it is this. And it is measurable: what percentage of the client’s revenue passes through the pipeline today, against the structural ceiling of how much could. The gap between those two is the actual work.
Organ, not cancer (complexity). Every new part must produce more value than it costs to maintain, and it must stay invisible to the host. The shape of the tree is the diagnosis: a system that widens is an organ, a system that deepens into custom, per-client machinery is a cancer.
Offer over client. The offer is the constant; the client fits into it. Remodeling the offer for each client is either the test phase or the cancer returning. This is the law that keeps the system replicable.
What keeps it from becoming a straitjacket
The trap in any framework is Procrustes: forcing every client onto the same bed. The escape is separating canon from instance. The chassis is canon: the strategic-to-tooling cascade, the grammar of the pieces, the sensor categories, the laws above. The biology is instance: this client’s specific focuses, engines, and teams. Same canon, different biology. That split is what lets you install the same operating system in two very different businesses without flattening either one.
One discipline holds it together: every strategic line has to terminate in a concrete task or a tool that reaches the client. If a line of thinking does not end in something that ships, it is a speech, and you cut it.
Takeaway
Automation amplifies whatever it points at, so the model is not the automation. It is the governance that makes speed safe. Separate the canon from the instance and the three laws stop being opinions: they become a filter that tells you, for any proposed piece, whether it is an organ worth installing or a cancer worth refusing.