Solution 07
A deterministic engine, not a black box.
The engine that defines your machine does not use generative AI. The same inputs produce the same outputs and a signed record you can reproduce later.
The problem
Engineering generated by AI is unsafe if you cannot check how it reached the answer. If the same prompt gives a different result tomorrow, the output cannot stand up to a certifier, auditor or customer.
The EU AI Act places high-risk AI systems, including AI in the safety components of products, under mandatory logging, traceability and human-oversight obligations. Read more
How Mitoma Systems solves it
The result comes from fixed engineering rules.
The engine does not use generative AI
AI interprets the mission and reviews the result. Fixed calculations and engineering rules define the machine. The engine cannot invent a tolerance or part because it is not guessing.
A signed trace, reproducible for years
Every decision records its inputs, the rule that fired and the result, hashed and signed. Feed the same inputs in a decade and you get the same machine, and the same explanation for why.
AI proposes, the engine decides, you audit
AI can suggest options and question results. Fixed engineering rules make the final technical decisions, and a qualified person approves them.
The output can face a certifier, an auditor or a customer, because every line of it traces to a rule and a source, and it produces the same answer every time.
Proof in the product
Reproducible by design, auditable by default.
The engine emits a signed decision trace with every artifact: the inputs, the rules that fired and the result. Any claim can be re-derived and checked long after the work is done.