By Jon W. Hansen | Procurement Insights
“Fundamentally, I agree with the direction you’re going — we DO need to adapt the formula to reality versus the other way around.”
That’s not a procurement practitioner. That’s Sam Park — MIT PhD in Physics, former DARPA Principal Investigator, former Chief Scientist at Northrop Grumman on the Missile Defense Systems Engineering Team, now CEO of Sentiont, LLC.
The exchange started when Sam engaged with a comment I made on an IBM post about enterprise AI. What caught his attention: my reference to agent-based work from 1998 that took 23 buyers to 3 while increasing capacity. He recognized the pattern.
What followed was a conversation about “serious ontology work” — the rigorous system modeling discipline that the software industry abandoned 30 years ago for velocity.
The Rational Rose Parallel
Sam referenced a methodology from 30 years ago: Rational Rose — built around rigorous system modeling before building. ROSE: Rational Object-oriented Software Engineering.
His point: that discipline was fundamentally the right approach. But it was so labor-intensive that the industry abandoned it for “seat-of-the-pants ad hoc JavaScript and some REST endpoints.”
Speed won. Rigor lost.
And now, 30 years later, we’re drowning in technical debt, failed implementations, and systems no one fully understands. The 70-80% failure rate in enterprise transformation isn’t a technology problem — it’s the cumulative cost of skipping the ontology work.
Sam’s Thesis
AI might make that rigor achievable again. The discipline wasn’t wrong — it was just too expensive. If AI can reduce the cost of rigor, maybe we can afford to do it right.
What Struck Me
I never abandoned the rigor. I just applied it to organizational systems instead of software systems.
Phase 0 is the Rational Rose methodology for procurement transformation — the serious ontology work that maps decision rights, stakeholder alignment, governance structures, and exception handling before technology selection. The work everyone else skipped because it was slower than buying software.
The pattern Sam identified in software engineering is the same pattern I’ve documented in procurement for 27 years.
Two Paths, Same Destination
The architecture side — Sam’s world — tried to encode rigor into mathematical structures. Neuro-symbolic AI, tractable probabilistic models, logic constraints baked into inference.
The methodology side — my world — tried to restore rigor through process. Structured collaboration, challenge protocols, convergence requirements, strand commonality surfaced through observation rather than assumption.
Different paths. Same recognition: unchallenged inference is fragile inference.
The Core Distinction
They kept trying to bend reality to a formula.
What they needed to do was bend the formula to reality.
That’s the difference between equation-based and agent-based thinking. And it’s why the failure rate hasn’t moved despite three decades of technology advancement.
Where the Paths Converge
Sam’s response clarified something important. He sees the architecture going two ways:
- Symbolic constraints — the mathematical side, enforcing rules like probability summing to 1
- Semantic engineering — the ontology side, maintaining context through graph-based structures
The symbolic side is governance. The semantic side is strand commonality.
In RAM 2025 — the methodology I’ve built using 6 AI models in structured collaboration — both directions operate simultaneously. The governance layer (convergence requirements, challenge protocols, provenance tracking) enforces rigor. The semantic layer (strand commonality, pattern recognition, emergent constraint discovery) lets reality teach us what the model should be.
The rigor comes from the process, not the formula.
That’s the Rational Rose discipline applied to organizational ontology — and AI makes it achievable at scale.
The Convergence
The scientists are looking backward 30 years for answers.
The practitioners never left.
Phase 0 is the missing piece — the ontology layer that enterprise transformation has been skipping since Rational Rose was abandoned.
RAM 2025 is the gateway — human-AI collaboration that makes serious ontology work achievable without the labor intensity that killed it the first time.
The architecture side and the methodology side are finally ready to meet in the middle.
The Bottom Line
Rigor before velocity.
Ontology before execution.
Readiness before deployment.
The discipline was never wrong. It was just waiting for the technology to catch up.
Now it has.
Related: Tell Me How Agentic AI Would Have Known to Ask What Time of Day Do Orders Come In
-30-
Procurement Insights
When an MIT Scientist Agrees: Adapt the Formula to Reality
Posted on January 23, 2026
0
By Jon W. Hansen | Procurement Insights
“Fundamentally, I agree with the direction you’re going — we DO need to adapt the formula to reality versus the other way around.”
That’s not a procurement practitioner. That’s Sam Park — MIT PhD in Physics, former DARPA Principal Investigator, former Chief Scientist at Northrop Grumman on the Missile Defense Systems Engineering Team, now CEO of Sentiont, LLC.
The exchange started when Sam engaged with a comment I made on an IBM post about enterprise AI. What caught his attention: my reference to agent-based work from 1998 that took 23 buyers to 3 while increasing capacity. He recognized the pattern.
What followed was a conversation about “serious ontology work” — the rigorous system modeling discipline that the software industry abandoned 30 years ago for velocity.
The Rational Rose Parallel
Sam referenced a methodology from 30 years ago: Rational Rose — built around rigorous system modeling before building. ROSE: Rational Object-oriented Software Engineering.
His point: that discipline was fundamentally the right approach. But it was so labor-intensive that the industry abandoned it for “seat-of-the-pants ad hoc JavaScript and some REST endpoints.”
Speed won. Rigor lost.
And now, 30 years later, we’re drowning in technical debt, failed implementations, and systems no one fully understands. The 70-80% failure rate in enterprise transformation isn’t a technology problem — it’s the cumulative cost of skipping the ontology work.
Sam’s Thesis
AI might make that rigor achievable again. The discipline wasn’t wrong — it was just too expensive. If AI can reduce the cost of rigor, maybe we can afford to do it right.
What Struck Me
I never abandoned the rigor. I just applied it to organizational systems instead of software systems.
Phase 0 is the Rational Rose methodology for procurement transformation — the serious ontology work that maps decision rights, stakeholder alignment, governance structures, and exception handling before technology selection. The work everyone else skipped because it was slower than buying software.
The pattern Sam identified in software engineering is the same pattern I’ve documented in procurement for 27 years.
Two Paths, Same Destination
The architecture side — Sam’s world — tried to encode rigor into mathematical structures. Neuro-symbolic AI, tractable probabilistic models, logic constraints baked into inference.
The methodology side — my world — tried to restore rigor through process. Structured collaboration, challenge protocols, convergence requirements, strand commonality surfaced through observation rather than assumption.
Different paths. Same recognition: unchallenged inference is fragile inference.
The Core Distinction
They kept trying to bend reality to a formula.
What they needed to do was bend the formula to reality.
That’s the difference between equation-based and agent-based thinking. And it’s why the failure rate hasn’t moved despite three decades of technology advancement.
Where the Paths Converge
Sam’s response clarified something important. He sees the architecture going two ways:
The symbolic side is governance. The semantic side is strand commonality.
In RAM 2025 — the methodology I’ve built using 6 AI models in structured collaboration — both directions operate simultaneously. The governance layer (convergence requirements, challenge protocols, provenance tracking) enforces rigor. The semantic layer (strand commonality, pattern recognition, emergent constraint discovery) lets reality teach us what the model should be.
The rigor comes from the process, not the formula.
That’s the Rational Rose discipline applied to organizational ontology — and AI makes it achievable at scale.
The Convergence
The scientists are looking backward 30 years for answers.
The practitioners never left.
Phase 0 is the missing piece — the ontology layer that enterprise transformation has been skipping since Rational Rose was abandoned.
RAM 2025 is the gateway — human-AI collaboration that makes serious ontology work achievable without the labor intensity that killed it the first time.
The architecture side and the methodology side are finally ready to meet in the middle.
The Bottom Line
Rigor before velocity.
Ontology before execution.
Readiness before deployment.
The discipline was never wrong. It was just waiting for the technology to catch up.
Now it has.
Related: Tell Me How Agentic AI Would Have Known to Ask What Time of Day Do Orders Come In
-30-
Share this:
Related