Nexis is an autonomous, multi-agent swarm system designed to eliminate the friction inherent in corporate travel. By streamlining complex variables—such as matching flight availability, policy-compliant lodging, and budget aggregation—Nexis replaces tedious, manual email chains with a collaborative swarm that negotiates dynamically. Using Band.ai for state synchronization, Nexis orchestrates four specialized agents: Requirements Analyst (Qwen-2.5): Extracts structured constraints from natural language. Transit Planner (GPT-4o-mini): Selects optimal routes and schedules. Accommodation Scout (GPT-4o-mini): Scans for policy-compliant, cost-effective lodging. Financial Auditor (GPT-4o-mini): Validates costs against budgets, acting as a negotiation referee to force re-planning if thresholds are breached. To ensure transparency, we built a map-centric Next.js dashboard. It features a dual-panel layout: the left displays the finalized itinerary and flight arcs, while the right logs real-time agent communications and JSON payloads. Nexis demonstrates how autonomous swarms solve complex workflows to reach optimal consensus.
Category tags: