AI Agent-Based Organizational Modeling: Dynamic Workforce Digital Twin Simulation Through Multi-Engine Composition

Digital twin technology has transformed manufacturing, supply chain management, and infrastructure monitoring, yet its application to organisations as complex adaptive systems — the interaction of people, processes, decisions, and financial constraints that determine enterprise performance — remains nascent. Existing approaches require either painstaking manual specification of agent archetypes (traditional ABM), lack financial grounding (organisational simulation), or demand direct ERP system access (vendor digital twins). This paper presents ABOM (AI Agent-Based Organizational Modeling), a novel approach in which the agent workforce is dynamically generated from the outputs of independent analytical engines rather than manually defined by domain experts. ABOM combines outputs from five validated engines within the Ingentic AI paradigm (Bhartiya, 2026a): Quarterly Process Intelligence (QPI) for financial constraints and regime classification, Simulated Process Mining for process flow templates, Industry Value Case for role archetype generation, Target Operating Model assessment for governance structure, and a Consulting Playbook for intervention trigger definitions. Given a company's public financial data, industry classification, and approximate employee count, ABOM automatically composes a hierarchically compressed workforce — representing thousands of employees through approximately 150 agents in five layers (C-Suite, VP/Directors, Department Heads, Specialists, and Aggregate Worker Pools) — that operates under quarterly financial constraints derived from actual KPI data. The paper introduces three innovations: (1) engine-driven agent generation, where workforce composition emerges from analytical engine outputs rather than manual specification; (2) hierarchical compression with aggregate worker pools, a novel agent type that models entire worker groups as statistical entities with headcount, utilisation, and capacity metrics; and (3) three simulation modes — Full-Human (baseline), Full-Agent (research experiment), and Hybrid Human+AI (the consulting deliverable, enabling evidence-based assessment of which roles to augment with AI and the projected financial impact). ABOM extends the Ingentic AI paradigm from backward-looking classification to forward-looking organisational simulation, completing the analytical cycle: Classification → Explanation → Prediction → Intervention.