Tegar Sang Putra

OpsMind AI is an evidence-based investigation assistant for manufacturing operations. It helps managers answer questions such as: “Why did overtime costs increase while production fell?” Relevant records are often spread across attendance, leave, overtime, production, and maintenance data, making investigations slow and easy to misinterpret. OpsMind detects anomalies using deterministic Go calculations, then uses Meta’s Muse Spark through the Meta Model API to select read-only tools and gather evidence across departments. It produces evidence-linked hypotheses, recommendations, counter-evidence, and information gaps. Managers can inspect source records, review recommendations, and record decisions in an audit trail. The system distinguishes unverified attendance from confirmed absence, excludes approved leave from misconduct assumptions, and treats correlations as hypotheses requiring verification. Human approval records a review decision without changing operational data. The bilingual prototype uses synthetic manufacturing data and has been validated with real Meta API investigations. Its goal is to reduce manual reconciliation and give managers a traceable basis for operational decisions.