FlexVertex

RecallIQ is a working AI-powered investigation and decision-reconstruction system built with Meta Muse Spark 1.3 and FlexVertex. It addresses a fundamental challenge in enterprise AI: how can an AI system reason over evolving evidence without losing track of what was known—and decided—at an earlier point in time? Using a synthetic food-safety investigation, RecallIQ reconstructs historical evidence graphs from FlexVertex Iron. As new laboratory findings and supplier corrections emerge, Meta Muse identifies affected production batches and explains how the investigation's scope changes. Deterministic validation ensures that conclusions remain grounded in the correct historical evidence. RecallIQ then persists Muse's bounded judgment, supporting evidence, model identity, rationale, and execution timestamps. A separate process can independently reconstruct and verify that judgment without calling the AI model again. The demonstration includes live database integration, Meta API inference, persistent judgments, independent replay, and MCP-based graph access. Although demonstrated through food safety, RecallIQ establishes a reusable architecture for auditable, historically accountable AI decisions across regulated industries.