Skip to content
English
Contact us
  • There are no suggestions because the search field is empty.

Problem Statement: Dispatch Alerts

💁‍♂️ Problem

Job-To-Be-Done (JTBD): "As a dispatcher, I need to be immediately alerted to operational exceptions and potential delays, so I can proactively manage my fleet and prevent customer-facing issues."

Pains to Resolve:

  • Reactive Operations: Dispatchers discover problems (like a late delivery) only after they've already occurred. They are constantly "fighting fires" instead of preventing them.

  • Information Overload: They must manually scan dashboards of raw data (lists of orders, maps of vehicles) to spot trends or anomalies. It's easy to miss a critical event.

  • Delayed Decision Making: The data they use for monitoring is minutes old, leading to poor and untimely operational decisions.

  • Inefficiency: Significant time is wasted investigating and responding to issues that a proactive alert could have mitigated or solved entirely.

🎯 Measuring Success

Leading Indicators (System Performance):

  • Alert Latency: Time from an exception condition being met (e.g., vehicle stopped too long) to the alert appearing on the dispatcher's screen is less than 5 seconds ($P_{99} < 5000ms$).

  • Alert Accuracy: Achieve a low false-positive rate for alerts to ensure dispatchers trust and act on the information.

  • Dispatcher Acknowledgement Time: Reduction in the average time it takes for a dispatcher to acknowledge and begin actioning a critical alert after it is issued.

Lagging Indicators (Business & User Impact):

  • Reduction in Service Level Agreement (SLA) Failures: Decrease in the percentage of late deliveries/orders by a target of 15% in the first six months.

  • Dispatcher Efficiency: Increase in the number of orders or assets a single dispatcher can effectively manage.

  • Reduction in Customer Complaints: Measurable decrease in customer-initiated complaints related to delays and exceptions.

🪚 Existing Solutions

  • Current Legacy Platform: Dispatchers use dashboards that auto-refresh on a multi-minute interval. Any "alerting" is done via end-of-day reports or by visually noticing a static indicator has changed to red. There is no real-time, proactive notification system.

  • User Workarounds: Dispatchers rely on experience, intuition, and direct phone calls with drivers to get ahead of problems. This is inconsistent, unscalable, and prone to human error.

  • Third-Party Tools: Some dispatchers may be using separate, disconnected tools (like fleet management software) that have better alerting, but this data is not integrated with our core order management system.

🤔 Proposed Solution

We will build an intelligent alerting feature on top of the new event-driven core platform. This solution moves beyond just showing data to providing actionable insights.

  1. Leverage the Event Bus: The real-time streams of data (GPS, status changes) already flowing through our event bus are the raw ingredients for this system.

  2. Complex Event Processing (CEP) / Rules Engine:

    • Implement a stream processing service (e.g., using Apache Flink, ksqlDB, or a custom microservice) that subscribes to multiple event streams.

    • This service will apply business logic and pre-defined rules in real-time to detect complex patterns that signify an exception.

    • Examples of Rules:

      • Trending Late: If a vehicle's current location and real-time traffic data predict an arrival time after the scheduled window, publish a TrendingLate event.

      • Unplanned Stop: If a vehicle's GPS reports a speed of 0 for more than X minutes while its status is In-Transit, publish a PotentialDelay event.

  3. Alert Generation & Delivery:

    • A new "Dispatcher Alerting Service" subscribes to these high-level exception events (e.g., TrendingLate, PotentialDelay).

    • When an exception event is received, this service pushes a formatted, actionable alert to the dispatcher's dashboard via the already-established WebSocket connection.

    • The UI will feature a dedicated alerts panel with visual and audible cues to draw immediate attention to the problem, along with deep links to the affected order or asset.

This solution transforms the dispatcher's role from a passive monitor to a proactive manager, allowing them to solve problems before they impact the customer.