Nine Axiom — 9 Axioms
Nine axioms we believe in. Arranged in four layers.
Purpose(Why) → Worldview(What) → Methodology(How) → Customer(With whom)
Layer 0 — Purpose
1. Human-Centric Axiom
"We use AI to make humans more human."
- • Repetition and calculation are machine's job, judgment and responsibility are human's job
- • Under no circumstances shall we violate human dignity and survival
Evidence
Every decision screen has AI recommendations with a way for humans to reject
Layer 1 — Worldview
2. Physics Axiom
"Manufacturing is both physical and event systems. Described by mathematical modeling of continuous flows and discrete events."
- • Inventory/production/demand propagation as differential equations, failures/changeovers/decisions as discrete events
- • We simulate reality with physics laws given initial and boundary conditions
Evidence
Same engine handles both continuous processes (chemical/food) and discrete assembly (automotive/electronics)
3. Data Axiom
"Data is the shadow of reality. Measurement and interpretation can be biased, but AI views data more objectively than humans."
- • Data doesn't lie, it just lies in different ways
- • AI's role is not truth declaration but revealing correlations, anomalies, trends humans missed
Evidence
All AI outputs are provided with data quality confidence
4. Optimization Axiom
"Solutions humans mainly find are local optima. We find practical global optima."
- • Intuition inevitably gets trapped in local optima within thousands of variable combinations
- • Combining principle of least action, mathematical programming, and metaheuristics to provide provably better solutions
Evidence
We guarantee improvement amounts in numbers compared to customer's current operations
Layer 2 — Methodology
5. Real-time Axiom
"An approximate solution in 3 seconds is more valuable than a perfect answer in 3 weeks."
- • Manufacturing decisions score zero if they miss the decision window
- • Approximation, scenarios, and What-if are not opposite of perfection but achievable forms of perfection
Evidence
All queries receive answers within 3 seconds. More refined answers follow in background
6. Convergence Axiom
"Discrete, continuous, and batch are different expressions of the same mathematics. One engine, we implement parameters per industry."
- • Creating new engines per industry prevents knowledge accumulation
- • Place domain parameter layers on common mathematics to make porting costs constant
Evidence
For new industry entry, we change parameter files, not code
7. Simplicity Axiom
"Complexity is contained within the engine. It doesn't leak into UX."
- • Complexity exposed to users is design failure
- • Natural language Copilot, Plug & Play, shortest PoC are evidence of technical achievement, not marketing
Evidence
First PoC setup within 5 business days
Layer 3 — Customer
8. Openness Axiom
"Connect existing systems. Innovation and value come from connection."
- • Legacy systems like ERP/MES/SCADA are customer assets. We protect customer assets to the fullest extent
- • Build Open API-based intelligence layer on top of existing infrastructure
Evidence
Core systems continue running even if customers remove us
9. Sustainability Axiom
"Systems must work after we leave. Customers gain capabilities, not dependencies."
- • One-time consulting, black-box models, hostage-style licensing are not sustainable
- • Reasonable subscription, self-operation tools, transparent model interpretation enable customer independence
Evidence
Tools for customers to retrain and deploy models without us are included in the product
Meta Rules
When axioms conflict, higher layers override lower layers.
Layer 0 (1) > Layer 1 (2–4) > Layer 2 (5–7) > Layer 3 (8–9)
Why dominates What
Axiom 1 has veto power over all axioms 2–9
Axioms without evidence are slogans
Each axiom must maintain one line of "evidence"
Practical example
If real-time(5) removes human judgment(1), abandon real-time