As-built redline processing that turns the crew's pen into checked GIS edits, so the network map matches the street in days, not months
From paper redline to GIS edit in days, not months, every edit traced to the pen.
See one case, screen by screen ↓When a crew finishes a job, the foreman marks up the construction print in red pen: the cable went down the other side of the road, the transformer sits 14 m further east, a link pillar went in that is not on the design. The print goes into a tray at the depot, then to the GIS team, where a technician redraws every mark by hand. At Lumina Utilities 1,240 sheets wait in that tray, 46 days on average from energisation to the map; the oldest is 214 days.
Meanwhile the control room switches from a map that does not show the new cable, planners design on stale network records, and handwriting like “way 3 or way 5” gets guessed. Completed work orders cannot close until the as-built is posted and the materials reconcile, so finished, paid-for plant sits in construction work in progress — $46.2M across 612 work orders at Lumina.
Estimated split of Lumina's 46-day average from energisation to the posted map, in days, by hand and with the solution.
Six agents lay the redline over the map, read the pen and draft each edit; the GIS technician accepts and posts.
Glen Mabry's crew energised the Harrow Lane 11 kV extension at Fernhill Estate phase 2 on Friday 9 October 2026 — work order WO-24-18837, feeder F-212 from Kestrel Road primary. Sheet 2 of 3 of his construction print was scanned at Westbrook depot at 07:52 on Tuesday. Here is what happens next, screen by screen, in the working solution.
Tomas posts network records for Westbrook. His tray holds 1,240 sheets, 46 days on average from energisation to the map; 113 are over 120 days and the oldest, Quayside, waits 214 days for a clearer scan. The agents read each sheet the moment it is scanned: 1,186 are ready for a technician, 31 need a clearer scan and 23 wait for a missing sheet. Harrow Lane, energised 4 days ago, is first.
“Today: 1,240 sheets, 46 days on average from energisation to the map; the oldest, Quayside, is 214 days.”
The registrar aligns the 400 dpi scan on 4 grid ticks, the Harrow Lane kerb lines and RMU-118, with an error of 0.6 m. The reader finds 11 marks, 10 read above 0.75. The drafter turns them into 9 proposed edits, every pen line snapped within 2.0 m. The design comparer finds 3 changes from design DP-24-0912 rev C, and the closeout tracker sees sheet 3 of 3 is not scanned yet and reminds Glen at 09:00.
Redline reader: “RMU way reads 3 or 5 (0.61), so you decide”
Glen's red ink sits exactly over the street: the designed south-verge route crossed out and a new one drawn along the north verge, “ran N side — 186m”; “TX moved 14m E”; “JT @ 92m”; “LP 2290 NEW”; the overhead span removed and pole P-7711 recovered. A fader moves between paper and map, and each mark carries a number that matches an edit in the list beside the map.
Glen's note on the sheet: “c/o gas main conflict – ran N side”
The sheet view shows every mark the agents found, what they read and their confidence: the joint at 0.88, the spare ducts at 0.81, the transformer move at 0.97. The reason note and Glen's signature and date, 09/10/26, are read too and matched to WO-24-18837. Below 0.75 a person decides.
“11 marks · 9 edits · 1 reason note · 1 signature”
Edit 1 proposes a primary 11 kV cable, 3×1c 300 mm² Al XLPE, 186 m against 172 m designed, from RMU-118 to TX-4471 on the north verge at 0.9 m cover. The route and length differ from the design but sit inside the +10 % rule with a reason on the sheet, so they are recorded. Tomas accepts, and the line snaps onto the verge and joins the two assets it meets.
“Edit 1 accepted — the cable snaps to RMU-118 and TX-4471”
Glen circled RMU-118 and wrote “fed way” with a digit that reads as 3 or 5. At 0.61 it is below the line, so it is Tomas's call. Switching schedule SS-F212 shows way 3 spare and planned for TX-4471, while way 5 already feeds the Teasel Close ring; recording way 5 would show two feeds on one way and fail the trace. Tomas picks way 3, then accepts the checked edits in one go.
“Way 3 — matches the switching schedule; edit 9 accepted”
Design vs as-built sets the design package, the crew's redline and the work order's materials side by side: 8 items compared, 5 match, 2 field changes recorded inside the rules, and link pillar LP-2290 is not on the design. The design comparer drafts the question; Tomas sends it. The materials agree with the pen — 200 m drum issued, 186 m laid, 14 m returned.
Adaeze Okafor: “Approved. We moved plot 23 off the overhead span, so the pillar replaces the pole service. I'll issue rev D for the record.”
The posting package holds the 9 accepted edits as one version, WO-24-18837-AB. The feeder trace runs F-212 from Kestrel Road primary through RMU-118 way 3 and the 186 m cable to TX-4471; the Teasel Close ring is unchanged; no orphans, no duplicates with sheet 1, and 9 of 9 features carry asset id, size, install date, crew and source mark. Tomas signs with a statement.
“I checked each edit against the crew's redline and accept them as built.”
Version WO-24-18837-AB is on the network map and the control room is told: a new 11 kV cable RMU-118 way 3 to TX-4471 (500 kVA) on the north verge, energised 9 October, new link pillar LP-2290, the overhead span P-7710 – P-7711 removed. Switching diagrams for F-212 update at posting. Redline to map: 4 days after energisation.
“Map change · feeder F-212 · Westbrook” — to the control room (Ravi Menon on shift), the outage desk and Westbrook planning
WO-24-18837, $412,600 of construction work in progress, moves from “Redline in · map not posted” to “Map posted · materials to reconcile”, with sheet 3 of 3 still awaited. The closeout tracker shows why the rest wait: 87 built with no redline scanned, 447 read and waiting for a technician, 31 design questions open, 29 contractor packages returned, 17 scans too poor to read.
Closeout tracker: never closes a work order or capitalises cost — Hannah closes it.
The same redline, seen by the five people who touch it — the crew, the GIS technician, design, closeout and the control room.
Six agents lay the redline over the map, read the pen and draft each edit; the GIS technician accepts and posts.
Reads handwriting, dimensions and hand-drawn symbols on scanned and photographed redlines — crossed-out routes, new routes, arrows, joints, link pillars, notes and signatures — each with a confidence, using Lumina's distribution and water symbol standards.
Aligns each scan to the network map from the print's grid ticks and coordinates, the kerb lines, street names and existing assets, and reports the registration error.
Turns each mark into one proposed edit — add, move, remove or change an attribute — with geometry snapped within 2.0 m to the verge, asset or duct it meets, and attributes filled from the design package, the work order and the stores issues.
Compares each edit with the design package and the work order's issued, used and returned materials, records field changes inside the standing rules, and drafts the question to the design engineer for anything new.
Traces feeders, phases and ring-main ways against the switching schedules, and finds orphans, duplicates and breaks before a version posts; sends the control room its map-change notice.
Follows each work order from built to closed, and sends reminders for missing sheets and contractor packages to the crew foreman and the closeout coordinator.
Accepts each edit; the design engineer approves new assets. The agents propose; a named person decides.
Ask about any mark, edit or sheet in plain words — or say what to change. Every answer points to the sheet, the design or the schedule behind it.
Why did the crew move the cable to the north verge?
Glen Mabry's note on the sheet says “c/o gas main conflict – ran N side”. Corran Gas's records put their 180 mm main in the south footway, right in the designed route. The north-verge route is 186 m against 172 m designed, inside the +10 % rule, so it is recorded without approval.
Is RMU-118 way 3 or way 5?
The digit reads as either, at 0.61 — below the 0.75 line, so it is yours to decide. Switching schedule SS-F212 shows way 3 spare and planned for TX-4471; way 5 already feeds the Teasel Close ring. Way 5 would fail the feeder trace. I suggest way 3.
What changed from the design?
Three things on sheet 2: the cable on the north verge, 186 m — recorded, gas conflict noted; TX-4471 14 m east — recorded, easement E-3317; and the new link pillar LP-2290, which needs Adaeze Okafor. The other five items match the design and the materials issued.
Which sheets have waited longest in the tray?
113 sheets are over 120 days. The oldest is WO-23-15288 Quayside, 214 days, waiting for a rescan because the right half is water-damaged. Next is the Riverside 33 kV diversion, 133 days, with 15 edits ready to accept.
12 screens from the working solution, on its sample data. Pick one to see it large.
How sheets are read, which changes need the design engineer and how fast the map must follow the street are settings, chosen by the GIS and records manager.
| Setting | Default | Choose from |
|---|---|---|
| A person decides below this confidence | 0.75 | Up or down in steps of 0.05 — readings under it are never proposed as ready to accept |
| Symbol library | Both standards | Lumina distribution standard 2024 · Lumina water standard 2023 · Both standards |
| Registration error allowed | 1.5 m | 1.0 m · 1.5 m · 3.0 m — a sheet that cannot be aligned this closely goes to a person |
| Snap to the network within | 2.0 m | 1.0 m · 2.0 m · 3.0 m |
| Recorded without approval | Length within +10 %, reason noted | Length within +10 %, reason noted · Length within +5 %, reason noted · Every change needs approval |
| Design engineer for Westbrook | Adaeze Okafor | Adaeze Okafor · Martin Hale — receives design-change questions |
| Target: on the map after energisation | 5 working days | 3 working days · 5 working days · 10 working days — the tray ranks sheets by this clock |
| Remind about a missing sheet after | 2 working days | 1 working day · 2 working days · 5 working days — to the crew foreman and the closeout coordinator |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample data · “target” = the design goal, measured in the live solution · “estimated” = our estimate from Lumina's demo inputs (612 work orders at an average $75.5K; the lag falling from 46 to 5 days), and the split of the 46-day lag · “typical” = published figures (1Spatial, case study of an Australian distributor: a 30-day KPI to get as-builts into GIS, aiming for 10–15 days). Lumina Utilities, Corran Gas and every person, place, contractor and work order named on this page are fictional — characters and sample data in the working solution.
It is the work of getting a crew's marked-up construction print — the redline — onto the GIS network map as built. As-Built Redline Processing does it with six agents that lay the scanned sheet over the map, read the handwriting and hand-drawn symbols, and draft each map edit with its geometry. A GIS technician accepts each edit and posts.
The sheet registrar aligns the scan to the map from its grid ticks, kerb lines and existing assets. The redline reader reads every mark with a confidence, the edit drafter turns each one into a proposed edit snapped to the network, the design comparer checks it against the design and the materials issued, and the connectivity checker traces the feeder before anything posts.
Yes. It reads the scans and photos already waiting, not only the next job captured digitally. In the working solution the agents read each sheet the moment it is scanned: of 1,240 sheets in the tray, 1,186 are read and ready for a technician, and 31 poor scans and 23 missing sheets are flagged.
A reading below the confidence line — 0.75 by default — is never proposed as ready to accept; a person decides. In the sample case the ring-main way read 3 or 5 at 0.61; the switching schedule showed way 3 spare and planned for the new transformer, and the technician chose way 3.
The feeder trace, ring, orphan, duplicate and attribute checks run before a version posts, and no version posts with a failed feeder trace. When an energised change posts, the control room, the outage desk and planning get a map-change notice naming the feeder and assets. The target is every energised change on the map within 5 working days.
Yes. Contractor CAD and PDF packages are checked against your as-built specification — seal and signature, grid and control points, scale, valve offsets, joint chainage, test certificates — before they reach the tray. A failed check goes back with the reasons drafted for you; a clean one goes to the tray.
Yes. Agents propose and never post, never move an existing asset to fit a sheet, never approve a design change and never close a work order. A named GIS technician accepts each edit and signs the version, the design engineer approves new assets, and the closeout coordinator closes the work order.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your symbol standards, as-built specification, design-change rules and a sample of redlines from your tray — and it goes live once every quality gate has passed. We will walk you through it on your own sheets first.
Clean asset history from every closed job, with the parts in the photos booked.
1-min demoWork Package PreparationCrews arrive with everything they need.
1-min demoLocate Request ScreeningClears go back in minutes; every machine dig over a high-voltage cable gets a standby.
1-min demoWe'll run As-Built Redline Processing on a sample of the sheets already in your tray.