AptStock AI — Voice Inventory Copilot

AptStock AI is a voice-powered demand forecasting and inventory optimization copilot for multi-SKU retailers managing hundreds or thousands of products. Multi-SKU stores already have sales, product, and inventory data; the operational challenge is turning that volume into a clear answer: what needs attention now, why, and what should happen next? AptStock closes that decision loop: REAL STORE DATA → DEMAND FORECAST → INVENTORY OPTIMIZATION → MULTI-SKU PRIORITY → VOICE REASONING → CONTROLLED REORDER → CHANGE/CANCEL → INTERRUPTION SAFETY → STATE VERIFICATION. Its intelligence layer analyzes historical sales and inventory context to forecast demand, identify inventory risk, prioritize products, apply safety-stock logic, and produce actionable replenishment recommendations. Instead of searching spreadsheets, reports, and dashboards, operators ask naturally by voice and receive explanations grounded in store context. The voice layer is not just a conversational front end. It operates the inventory workflow. In the demonstrated flow, the agent identifies a priority item, prepares a reorder draft, accepts a correction to the requested quantity, preserves the updated state, and handles a cancellation request without silently placing the order. It then verifies the resulting state. This makes the interaction a controlled operational workflow rather than a voice-only demo. For the AssemblyAI Voice Agent Hackathon, AptStock uses AssemblyAI's Voice Agent API for real-time conversation, turn-taking, interruption handling, and tool calling. AptStock connects that voice experience to its forecasting, inventory, and action services so spoken requests can move from insight toward a controlled operational decision. The result is a practical bridge between predictive retail intelligence and store operations: AptStock determines what deserves attention, explains why it matters, and gives an operator a natural way to understand, change, cancel, and control the next action.