Telecommunications networks are evolving toward “connected intelligence,” where autonomous control loops must operate across dynamic radio environments, multi-vendor stacks, and stringent regulatory constraints. In this context, ATIS introduces NESY — a reference architecture that couples neural perception and prediction with symbolic reasoning and verification to deliver bounded autonomy with audit-ready decision traces. NESY is designed to augment, rather than replace, existing controllers; it correlates multimodal evidence, verifies candidate actions against explicit policies, and produces a unified Decision-Action-Explanation (DAE) output. By bridging the gap between AI pattern recognition and the need for accountability, NESY ensures that each automated decision is accompanied by a rationale that can be inspected by operators and regulators. Explore the full paper to understand how this architecture sets a foundation for trustworthy, transparent, and resilient AI in next generation networks.