Implementation Roadmap

Current State (Assumed)

Based on the data analysed, the current scheduling process is understood to follow a manual, periodic approach:

1. Extract — Due maintenance tasks are extracted from the ERP (e.g. Ellipse) based on upcoming due dates and tolerance windows
2. Spreadsheet planning — Work orders are distributed to work groups in Excel, with planned dates placed at the midpoint of each tolerance window
3. Monthly review — The plan is reviewed and adjusted periodically (likely monthly) to account for completions, deferrals, and changing priorities
4. Distribution — Work packs are issued to field teams, typically by work group section

This process treats each work order independently. There is no automated consideration of geographic proximity, team overlap, or tolerance window optimisation when assigning planned dates or work groups.

Proposed Future State

The proposed solution introduces Envision as the central planning system, replacing the manual spreadsheet-based approach with an automated, continuously optimised schedule.

Architecture

┌─────────────┐     Live Interface      ┌──────────────────┐
│             │ ───────────────────────> │                  │
│   ERP       │   Assets, Due Dates,    │    Envision      │
│  (Ellipse)  │   Tolerances, WOs       │  Planning Engine  │
│             │ <─────────────────────── │                  │
└─────────────┘   Optimised Schedule,   └──────────────────┘
                  Planned Dates              │
                                             │  Optimised
                                             │  Work Packs
                                             ▼
                                    ┌──────────────────┐
                                    │   Field Teams    │
                                    │  (Daily Plans)   │
                                    └──────────────────┘

What Envision Would Do

1. Ingest — Receive asset register, maintenance requirements (TMP), due dates, tolerance windows, and work order status from the ERP via a live interface
2. Optimise — Automatically generate an optimised schedule that:
3. Batches work at the same location within tolerance windows
4. Assigns work to teams based on depot proximity and capability
5. Sequences daily routes to minimise travel
6. Flags and resolves multi-team overlaps at boundary stations
7. Balances workload across teams and days
8. Publish — Push optimised planned dates and work group assignments back to the ERP
9. Adapt — Continuously re-optimise the forward schedule as work is completed, new tasks become due, or conditions change

Key Differences from Current Process

Implementation Phases

Phase 1: Foundation (Months 1-3)

Objective: Establish the data pipeline and prove the optimisation engine on live data.

Process change: Minimal. The ERP remains the system of record. Envision generates an optimised schedule that can be reviewed alongside the current Excel plan. No change to field operations.

Expected outcome: Demonstrate measurable reduction in planned visits at pilot depot.

Phase 2: Optimisation Engine (Months 3-6)

Objective: Build the full scheduling engine and pilot at one depot.

Process change: Moderate. The planning team reviews Envision's optimised schedule instead of building one from scratch in Excel. Field teams receive work packs generated from Envision. The ERP remains the system of record for work order status.

Expected outcome: Validated reduction in visits, travel, and overlaps at pilot depot. Planner time freed from manual schedule construction.

Phase 3: Full Deployment (Months 6-12)

Objective: Roll out across all depots and establish Envision as the primary planning system.

Process change: Significant. Envision becomes the central planning system. The monthly Excel planning cycle is replaced by a continuously optimised, live schedule. Planners shift from schedule construction to exception management and oversight.

Expected outcome: Network-wide efficiency gains. Reduced travel, fewer unnecessary mobilisations, better team utilisation, and improved TMP compliance through automated tolerance window management.

Phase 4: Continuous Improvement (Ongoing)

Investment Summary

Detailed scoping and costing for each phase would be provided as a separate proposal following acceptance of this POC.

Dependencies and Assumptions

1. ERP access — A live data interface (API or database view) to the ERP is required. The specific integration method depends on the ERP platform and IT architecture.
2. Asset coordinates — All maintainable assets need geocoded locations. This POC found 89% coverage in the source data, with the remainder geocoded from station names.
3. Job durations — Duration estimates per standard job are not in the current data. These would need to be sourced from the ERP or estimated from historical completion data.
4. Stakeholder engagement — The planning team needs to be involved in the design of the optimised workflow to ensure the system fits their operational reality.
5. Change management — Moving from a manual Excel process to an automated system requires training, parallel running, and clear escalation paths for exceptions.