User Story: System Alerts & Proactive Recommendations (AX03US03)

1. Problem Statement

This user story addresses the need to move beyond scheduled maintenance and reactively fixing failures by proactively identifying at-risk assets before they fail. The primary user is the Asset Manager, who is responsible for managing the long-term health and risk profile of the utility's entire asset base.

Primary User Role & Pain Points:

• Asset Manager:
  • Blind to Emerging Risks: Relies on fixed schedules (Preventive Maintenance) or waiting for failures (Corrective Maintenance), leaving them blind to assets that are degrading faster than expected between cycles.
  • Inability to Act Proactively: Lacks a system that can synthesize complex data (e.g., condition, criticality, operational data) to flag specific, high-risk assets that need immediate attention.
  • Difficulty Prioritizing "Hidden" Work: Struggles to decide which non-urgent work to perform. Without data-driven recommendations, they may service a low-risk asset while a high-risk one is silently approaching failure.
  • Missed Cost-Saving Opportunities: Cannot quantify the potential savings of performing proactive maintenance, making it difficult to justify the allocation of resources away from scheduled tasks.
  • Overwhelmed by Data: Has access to vast amounts of asset and operational data but lacks the tools to automatically analyze it and surface actionable insights.

Core Problem:
For the Asset Manager, traditional maintenance strategies are insufficient to prevent all critical failures. Without an intelligent system to analyze data and predict potential issues, they are perpetually one step behind, unable to identify and mitigate emerging risks proactively. This leads to preventable failures, higher emergency repair costs, and an inability to optimize maintenance spending effectively.

2. Who Is the User Facing the Problem?

The Asset Manager is the most important user for this feature. Their core mandate is to manage and mitigate risk across the asset lifecycle. This "System Alerts" module acts as their early-warning system, directly feeding them the data-driven, prioritized insights needed to make strategic decisions about where to intervene. While an O&M Manager would execute the resulting work, the Asset Manager is the one who reviews, validates, and approves these proactive recommendations as part of their overall risk management strategy.

Access Control:
The Asset Manager should have full access to review, action, and dismiss recommendations. The O&M Manager would also have full access to convert recommendations into Service Orders. The Utility Administrator would have configuration access.

3. Jobs To Be Done

• For the Asset Manager: When I need to prevent critical asset failures and optimize my maintenance budget, but I cannot see which assets are silently becoming high-risk between scheduled inspections, help me by providing an AI-powered system that analyzes asset data to generate prioritized maintenance recommendations, so that I can intervene before failures occur, reduce emergency repair costs, and allocate my resources to the most significant risks.

4. Solution

The proposed solution is a System Alerts module that functions as an AI-powered recommendation engine. It analyzes assets with low condition scores and high risk scores to generate "Proactive Maintenance Insights." This provides a prioritized list of potential interventions, complete with context and actions to convert insights into work.

Key Capability Areas:

1. Proactive Insights Dashboard:
   • High-level KPIs summarizing the state of proactive maintenance: Total Active Recommendations, High-Risk Assets Identified, Potential Cost Savings, and Recommendations Overdue.
2. Prioritized Recommendation List:
   • A central table listing all system-generated recommendations.
   • Each recommendation includes an ID, the associated Asset, Location, Condition Score, Risk Score, a description of the Potential Impact, and its current Status.
3. Risk-Based Prioritization:
   • The core logic of the feature is to identify assets with a combination of a low Condition Score (high probability of failure) and a high Risk Score (high consequence of failure).
   • This ensures the most important and vulnerable assets are surfaced first.
4. Actionable Workflow:
   • Each recommendation has a clear status (e.g., New, Reviewed, Action Taken, Overdue, Dismissed).
   • A simple "Actions" menu allows the Asset Manager to:
     • View Details for deeper analysis.
     • + Create Service Order to directly convert the recommendation into a work order.
     • Mark as Reviewed to acknowledge the insight.
     • Dismiss Recommendation to close it out with a justification.
5. Quantified Business Value:
   • The Potential Cost Savings KPI estimates the financial benefit of acting on these recommendations over a given period, providing a clear ROI for the feature.
   • The Potential Impact column for each recommendation describes the specific negative outcome that the proactive work would prevent.

5. Major Steps Involved

User Role: Asset Manager

Flow 1: Reviewing and Actioning a High-Risk Recommendation

1. Navigate to O&M -> System Alerts.
2. Review the KPI cards on the dashboard. Note there are 8 High-Risk Assets Identified.
3. Scan the "Proactive Maintenance Insights" table and sort by "Risk Score" in descending order.
4. The top item is INSIGHT-2025-001 for the "Primary Intake Pump," which has a low Condition Score (2.8) and a high Risk Score (8.2).
5. Read the Potential Impact: "Service disruption affecting 12,500 consumers."
6. Recognizing the severity, click the "..." (Actions) icon for that row.
7. From the dropdown menu, select "+ Create Service Order".
8. The system navigates to the "Create Service Order" screen, pre-filling it with information from the recommendation (e.g., Title, Asset selection, Priority set to 'Critical').
9. The Asset Manager completes any remaining details and creates the service order.
10. Upon returning to the System Alerts screen, the status of INSIGHT-2025-001 has automatically changed from "New" to "Action Taken".

6. Flow Diagram

7. Business Rules

This section provides a detailed breakdown of rules for every visible element on the screen.

A. System Alerts (Dashboard View)

• Page Title/Subtitle: Static text explaining the feature's purpose.
• KPI Card: "Total Active Recommendations"
  • Formula: COUNT(Recommendations) where Status IN ('New', 'Reviewed', 'Overdue').
• KPI Card: "High-Risk Assets Identified"
  • Formula: COUNT(Recommendations) where Risk Score >= 7.0 (or another defined threshold) AND Status IN ('New', 'Reviewed', 'Overdue').
• KPI Card: "Potential Cost Savings"
  • Formula: SUM(EstimatedCostOfFailure - EstimatedCostOfProactiveRepair) for all recommendations where Status = 'Action Taken' within the "last 30 days". This is a complex metric requiring predefined cost estimates.
• KPI Card: "Recommendations Overdue"
  • Formula: COUNT(Recommendations) where Status = 'Overdue'. A recommendation becomes overdue if Action Date has passed and status is still 'New' or 'Reviewed'.
• Search Bar: Must search across Recommendation ID, Asset ID/Name, and keywords in the Potential Impact description.
• Refresh/Filters Buttons: Standard controls to refresh the data and apply advanced filters (e.g., by Status, by Risk Score range).
• Table: "Proactive Maintenance Insights"
  • Recommendation ID: Unique, system-generated ID for the insight.
  • Asset ID / Name: The specific asset identified.
  • Location: The location of the asset.
  • Condition Score: The asset's current condition score. Must be color-coded (e.g., red for < 3.0).
  • Risk Score: The asset's current risk score. Must be color-coded (e.g., red for > 7.0).
  • Action Date: The suggested date by which action should be taken.
  • Potential Impact: A text description of the business consequence of failure.
  • Status: A colored tag indicating the recommendation's state (New, Reviewed, Overdue, Action Taken, Dismissed).
  • Reviewed By / Reviewed On: If status is 'Reviewed' or later, these fields are populated with the user's name and the timestamp.
  • Actions ("..."): A menu button that must reveal:
    • View Details: Opens a detailed view of the insight and the underlying data.
    • + Create Service Order: Initiates the workflow to create a new SO, pre-populating it with data.
    • Mark as Reviewed: Changes the status to 'Reviewed' and populates the 'Reviewed By/On' fields.
    • Dismiss Recommendation: Opens a prompt asking for a reason, then changes the status to 'Dismissed'.

8. Sample Data

Recommendation Record (for INSIGHT-2025-001):

• Recommendation ID: INSIGHT-2025-001
• Asset ID / Name: PUMP-001, Primary Intake Pump
• Location: Main Treatment Plant
• Condition Score: 2.8
• Risk Score: 8.2
• Action Date: Jan 15, 2025
• Potential Impact: Service disruption affecting 12,500 consumers
• Status: New
• Reviewed By: N/A
• Reviewed On: N/A

9. Acceptance Criteria

1. The system must generate recommendations for assets that cross a certain threshold (e.g., Condition Score < 4.0 AND Risk Score > 6.0).
2. The system must display the four KPI cards with correctly calculated values.
3. The system must display a searchable and filterable list of all active recommendations.
4. The system must color-code the Condition and Risk scores to draw attention to poor values.
5. The system must provide an "Actions" menu for each recommendation with four options: View Details, Create Service Order, Mark as Reviewed, Dismiss.
6. Selecting + Create Service Order must launch the SO creation wizard and pre-populate relevant data.
7. After a service order is created from a recommendation, its status must automatically update to "Action Taken".
8. Selecting Mark as Reviewed must update the status and log the user and timestamp.
9. Selecting Dismiss Recommendation must prompt for a reason before changing the status.
10. The Recommendations Overdue KPI must accurately count recommendations that have passed their Action Date without being actioned or dismissed.
11. The Potential Cost Savings KPI must correctly calculate the financial benefit based on its formula.
12. The system must allow filtering recommendations by their Status.
13. The user must be able to sort the recommendation list by any column, especially Condition Score and Risk Score.
14. The High-Risk Assets Identified KPI must use a configurable threshold (e.g., Risk Score >= 7.0).
15. Dismissed recommendations should be hidden from the default view but accessible via a filter.

10. Process Changes

11. Impact from Solving This Problem

12. User Behavior Tracking

Primary User Role: Asset Manager

Wireframe

Link