Causal Reasoning and Hybrid Intelligence in Mining Operations

17 Apr 2026

The Question the Industry Is Getting Wrong: Flowchart or LLM Is a False Choice

How causal reasoning and hybrid AI architecture solve the problem that neither structured procedures nor language models can solve alone

The NTT DATA and MIT Technology Review study on mining autonomy in Latin America (2025) identified something counterintuitive: the most important barrier to AI adoption in mining is not resistance to change or budget constraints — it is the design of the reasoning system itself. When an AI system can recommend the right action but cannot explain why in the language of the operator who must execute it, adoption fails. When it is too rigid to adapt to the variability of a real shift, it fails. When it cannot distinguish between what happened and what caused it, it fails late.

This whitepaper is the second pillar of NTT DATA’s Mining Autonomy Operating Model — The Intelligence layer. It introduces the hybrid reasoning architecture developed through NTT Research (published at ACL 2025) that powers the MOV operational intelligence platform: a dynamic causal graph that understands the temporal chains in which mining problems propagate, a three-mode reasoning engine that combines structured flowcharts with large language models based on available confidence, and a learning loop that makes the system wiser with every shift. The result is an intelligence layer that anticipates rather than reacts — and that the operator can trust because they can see how it reasons.

Download the Casual Reasoning and Hybrid Intelligence in Mining Operations Whitepaper

Name *

Last Name *

Position *

Company *

Industry *