.png&w=256&q=75)
1
1
Looking for experience!

FlowWatch is an AI Command Center for proactive telecom TV streaming QoE monitoring and incident response. Streaming customers can experience poor quality because of buffering, latency, packet loss, low bitrate, app crashes, or playback instability. These issues are often detected only after a customer complains. The app supports multiple telemetry sources: manual input, a built-in demo live API feed, external API telemetry, embedded HLS player telemetry, and live browser playback metrics. It calculates a QoE score, evaluates playback smoothness using a Playback Impact Gate, and classifies the session as stable, at risk, impact confirmed, or critical. FlowWatch uses a multi-agent workflow. The QoE Monitoring Agent evaluates streaming quality. The Incident Commander Agent decides whether to monitor, diagnose, self-heal, contact customer care, or escalate. The Diagnosis Agent identifies the likely root cause. The Recovery Action Agent recommends safe recovery actions. The Customer Care Agent prepares customer-safe communication and support summaries. A key feature is the Playback Impact Gate, which avoids false positives by checking whether playback is truly affected before agents are triggered. For confirmed issues, FlowWatch can recommend safe self-healing actions such as refreshing the streaming session. The operator can approve or reject the action, after which the app simulates recovery, recalculates QoE, and verifies improvement. Band is used as the incident collaboration layer. FlowWatch can publish agent decisions, task handoffs, approval requests, recovery recommendations, and audit traces into a Band incident room. FlowWatch demonstrates how AI agents can support telecom service assurance through telemetry monitoring, playback-aware decisions, autonomous orchestration, safe recovery, human approval, and Band-powered collaboration.
19 Jun 2026