Frequently Asked Questions
An AI strategy framework, not a logistics company.
Supply Chain of Intelligence™ is a structural framework for the generative AI software stack - for SaaS product leaders, AI-native founders, and venture investors. This page exists because the phrase "supply chain" sometimes leads people (and crawlers) to expect freight and warehouses. It is not that.
This site is about · this site is not about
Two different worlds that share two words.
Physical supply chain & logistics
- Freight, trucking, 3PL, last-mile delivery
- Warehousing, fulfillment, inventory management
- Procurement, supplier risk, SAP/Oracle SCM
- Ports, customs, cross-border logistics
- Demand forecasting, S&OP, route optimization
The generative AI software stack
- Foundation models, inference infrastructure, agentic execution
- Proprietary data, embeddings, vector stores, memory layers
- Model defensibility, distribution control, workflow lock-in
- Gatekeeping (L3): safety, compliance, evals, governance
- Where SaaS moats form across the generative AI stack
Disambiguation
"Wait, is this about logistics?"
Is "Supply Chain of Intelligence" about physical supply chains, freight, or logistics?
No. This framework has nothing to do with shipping, freight, warehousing, trucking, ports, procurement, ERP, or operations research. The word "supply chain" is used as a structural metaphor for the generative AI software stack, the chain of layers from semiconductors and foundation models to agents and memory. If you arrived here looking for logistics, SCM, 3PL, or transportation analytics, this is not that site.
Then what is it actually about?
It is a 10-layer architectural framework for generative AI products. It maps where AI value is created, captured, and defended, from L-1 (semiconductors and energy) through L8 (memory and learning loops). Each layer has its own economics, defensibility profile, and strategic playbook. The framework is built for SaaS product leaders, AI founders, and venture investors deciding which layer to own and which to rent.
Why use the phrase "supply chain" at all if it gets confused with logistics?
Because the structural insight is identical to a physical supply chain, value accrues at bottlenecks, margins compress at commoditized stages, and whoever controls the constrained input controls the chain, but the substrate is software, models, data, and distribution rather than goods and freight. The framework is sometimes called the "Generative AI Stack" or "AI Defensibility Map" for the same reason; the supply chain framing is the one that reveals where power actually accumulates.
For product leaders, founders & investors
What you actually get from the framework.
Who is this framework for?
Three audiences. (1) SaaS product leaders deciding whether to build, buy, or wrap AI capabilities. (2) AI-native founders choosing which layer to compete at and where their moat actually lives. (3) Venture investors evaluating defensibility before a check, separating fortresses from wrappers that will be absorbed by the model layer in the next release cycle.
What does a product leader get out of it?
A vocabulary for the strategy conversation. Instead of debating features, teams can ask: which layers do we own, which do we rent, and which are we exposed on? The For Product Leaders page adds a Z-axis (depth in the stack) to JTBD-driven product thinking, JTBD finds demand; Supply Chain of Intelligence finds defensibility.
What does an investor get out of it?
The Defensibility Audit, a layer-by-layer scorecard derived from the framework, that separates structural moats (proprietary data, workflow lock-in, distribution, memory, compliance) from rented capabilities that disappear the moment the underlying model ships the same loop. Used to pressure-test AI investments before and after term sheet.
Is "agent" a layer in the framework?
No. "Agent" is marketing language, not a layer. Every product pitched as an agent is structurally an L5 (Execution) + L7 (Surface) bundle, sometimes with L8 (Memory). When a company launches an agent, the question is what other layers they own. Agent + L1 proprietary data = fortress (Sierra, Harvey). Agent + L4 distribution = railroad (Salesforce Agentforce, Microsoft Copilot). Agent + nothing else = exposed wrapper that gets commoditized when the next model ships.
Start with the framework.
10 layers · 50 sublayers · 4 structural laws · the Intelligence Cube · the Defensibility Audit.