1.1 Reinforcing feedback loops These drivers do not act independently. A systems thinking lens reveals reinforcing feedback loops that sustain and amplify capital escalation: ● R1: Risk-driven overdesign: Uncertainty increases perceived project risk, which drives more conservative design choices and scope additions, further embedding capital as the response to risk. ● R2: Interface complexity: Additional scope creates more interfaces, which increases integration risk and reinforces the perceived need for further design margin. ● R3: Late confidence: When system behaviour is not tested early, teams carry assumptions into later project stages where changes become significantly more costly. ● R4: Cost escalation pressure: As capital intensity rises, scrutiny and risk sensitivity increase, which can paradoxically reinforce conservative decision-making and additional scope. The net effect is that the project system can unintentionally buy confidence through additional capital, rather than creating confidence through evidence. The central problem is not that individual scope additions are unreasonable, but that the system lacks the means to test their collective necessity before capital is committed. 2. OBJECTIVES AND SCOPE This paper presents an integrated framework for improving capital efficiency in complex mining projects by combining three complementary capabilities: ● Systems thinking to identify where and why capital is being added to compensate for uncertainty, shifting the conversation from defending scope to understanding the system outcomes that scope is intended to protect. ● Mining value chain digital twins to create a controlled environment in which alternative design options can be tested under realistic operating variability before capital is committed. ● Artificial intelligence and reinforcement learning to accelerate scenario discovery, identify patterns and constraints, and translate simulation outputs into decision-ready insights. A case study from a maintenance facility project demonstrates the practical application and outcomes. The paper also addresses the conditions under which this approach can scale across project portfolios. 3. METHODOLOGY AND APPROACH The framework integrates three layers, systems thinking, digital twin simulation and AI-
RkJQdWJsaXNoZXIy MTM0Mzk2