Work order closeout validation that reads every closed job, photos included, before it reaches the asset's history
Clean asset history from every closed job, with the parts in the photos booked.
See one case, screen by screen ↓When a crew closes a job from the van, what it leaves behind is the asset's history: what failed, why, what was done, which parts went in, how long it took and what else they saw. Most closeouts pass with “replaced fuse, supply restored”, a blank cause code, one person's hours, and a part that went into the asset from van stock but never onto the job. A planner who reviews 40 to 60 closeouts before planning the week cannot read every photo and timesheet, so most organisations sample and the rest pass unread.
The cost shows up later. A blank cause code is a failure the reliability engineer can't learn from: when the same failure is written five ways in five work orders, no query can group them, so there is no trustworthy Pareto or MTBF. Unbooked labour makes the next plan too short; unbooked parts understate job cost and leave stock drifting. And a defect written only in free text — “crossarm cracked, wants changing” — has no follow-up work order, so it comes back as an unplanned dispatch or the next outage.
Estimated split for reviewing one night's 46 closeouts by hand and with the solution, in hours. Our estimate, not a measured result.
Six specialist agents read every closed job, photos included; the planner accepts or sends it back, the crew confirms.
At 02:14 on Monday 12 October 2026, 23 properties on Larch Lane, Ashcombe lose supply. Crew N-14 — Dylan Marsh, crew lead, and Owen Pryce, lineworker — find a blown L2 LV fuse at pole-mounted transformer PMT T-48211 on pole AS-2207, restore supply at 04:40 and close the job from the van at 05:10. Here is what happens next, screen by screen, in the working solution.
Forty-six jobs closed since 18:00 on Sunday across Distribution North and Water South, all checked by 07:05. Each is a row with six marks — codes, time, parts, photos, readings, follow-up — and a quality score, worst first. 29 are clean and ready to accept in one click; 17 have 21 gaps between them. The Larch Lane fuse job sits at the top, scoring 58.
“46 closed since 18:00 Sun · 29 clean · 17 for you · 21 gaps”
Dylan's note reads: “L2 LV fuse blown at PMT. Bird strike on LV mains span to pole AS-2208, remains at base. Replaced fuse, supply restored 04:40, all 23 on. Crossarm at the cutouts is cracked, wants changing.” Failure mode ELV-02 and activity REP-01 are coded; the cause is blank. The margin notes tie each sentence to a gap: the cause, a fuse-link not on the parts list, a defect not raised — and one that matches the fault record.
“Cause is blank. The notes name a bird strike. Suggested cause EXT-03 External — wildlife contact.”
Photo 1: the L2 fuse-link operated. Photo 2: a new fuse-link, its label “100A gG 415V” legible — not on the parts list. Photo 3: the crossarm at the cutouts with a crack about 40 cm long. Photo 4: 239.4 V L2–N after restoration. Van V-114 issued one LV fuse-link 100 A gG at 03:58; the job's parts list has only four cable ties.
The photo reader never books a part. Below 0.85 confidence it asks rather than proposes.
The timesheets show Dylan Marsh and Owen Pryce on the job from 02:40 to 06:10 — 3.5 hours each. Only Dylan's 3.5 hours are booked. Against a plan of 4.0 hours for a standard LV fuse job, 7.0 hours worked is 75 % over plan, outside the ±25 % tolerance, so Grace Okafor, the work management supervisor, sees it too. The closeout summary is already written: issue, diagnosis, action and parts.
“Owen Pryce worked 3.5 h on this job and it is not booked.”
The run trace shows each agent's work. The closeout checker applied the rules for Corrective · LV fuse operation in 0.4 s. The notes reader matched “bird strike on LV mains” to EXT-03 at 0.93 confidence. The follow-up finder found no open work on pole AS-2207 and drafted “Replace LV crossarm”, P3, 2 × 3.0 h. The repeat failure watcher found three LV fuse operations on T-48211 in seven months.
“Score 58 / 100 · Tomás Reyes decides: accept, or send back with the gaps”
T-48211 is a 200 kVA, 11 kV / 415 V transformer feeding 23 properties, installed in May 2011. Its L2 fuse operated on 14 March and 2 August too, both closed before closeouts were checked: no cause, no parts. Today's closeout is the first to give one — a bird strike on the span to AS-2208 — and the watcher suggests a wildlife guard on that span, fitted with the crossarm change, for Henrik Solberg, the reliability engineer.
“Repeat failure · three LV fuse operations in 7 months”
Tomás adds the fuse-link for the crew to confirm: the line goes on the parts list as proposed, linked to photo 2 and the van V-114 issue. He accepts the suggested cause, EXT-03 External — wildlife contact. He approves the follow-up, and “Replace LV crossarm, pole AS-2207” is raised as WO 3318502, within 28 days.
“Follow-up raised · WO 3318502 · Replace LV crossarm, pole AS-2207”
Two items go to Dylan Marsh, crew lead, in the crew app and by text: book Owen Pryce's hours, +3.5 h, and confirm the LV fuse-link 100 A gG used. Each item goes with its evidence — the photo, the timesheet line, the note.
“When the crew replies the checks run again; nothing is booked until they confirm.”
On his phone Dylan sees the request from Tomás with the timesheet line for Owen's hours and photo 2 with the van issue. He taps Yes twice, and the job shows confirmed at 07:52. The checks run again on his reply and every gap on the job is closed.
“4 jobs sent back in the last 24 hours · 3 confirmed · median reply 47 minutes”
On planned vs actual, the fuse-link line now reads confirmed. Three more parts seen but not booked overnight are still open — a pair of 150 mm repair clamps, a spindle seal kit and four ABC connectors — making four lines and $412.80 with the fuse-link. The stock adjustment list goes to the Ashcombe and Fenwick stores. T-48211's history now reads: LV fuse operated · external — wildlife contact · replaced LV fuse-link 100 A gG · 7.0 h · follow-up WO 3318502 raised.
An estimated $262,392 a year of parts put back on the right job and asset — Lumina's own pilot inputs.
The same night fault, seen by the five people who rely on its closeout — what their week looked like, and what it looks like now.
Six specialist agents read every closed job, photos included; the planner accepts or sends it back, the crew confirms.
Applies the job type's closeout rules — failure mode, cause and activity codes, required photos and readings, timesheet match and parts against van stock — and scores every completed job.
Reads every crew photo: parts and their ratings from labels, before and after, visible damage, meter and gauge readings. Compares what was fitted with the parts list and proposes missing lines.
Turns the crew's notes into a structured summary — issue, diagnosis, action, parts — and suggests failure mode, cause and activity codes from the code standard, citing the sentence each comes from.
Matches timesheets, crew rosters and van-stock issues to the job, and compares planned with actual labour and materials against the tolerances.
Finds defects in the notes and photos, checks for open work on the asset and drafts the follow-up work order with a priority and an estimate.
Reads the asset's coded history and flags repeat failures — three operations in 12 months — to the reliability engineer.
Accepts or sends back; the crew confirms. The agents propose; a named person decides.
Ask about any closed job, gap, part or crew — or say what to add or send back. Nothing is written to a job until the planner or the crew confirms.
Why did the Larch Lane job score 58?
Three things are missing and one needs raising. Parts: photo 2 shows a new 100 A gG fuse-link and van V-114 issued one at 03:58, but the parts list has only cable ties. Time: Owen Pryce's 3.5 h is on the timesheet, not on the job. Cause: blank, though the notes say “bird strike on LV mains”. And the cracked crossarm in photo 3 has no work raised.
Add the 100 A fuse-link for the crew to confirm
Added 1 × LV fuse-link 100 A gG, $38.40, to WO 3318426, marked proposed — crew to confirm. It is linked to photo 2 and the van V-114 issue at 03:58. It goes to Dylan Marsh with the send-back; nothing is booked until he confirms.
Which transformers had three fuse operations this year?
Three, from the coded history. T-19033, Ferris Green: 4. T-48211, Larch Lane: 3 (14 Mar, 2 Aug, 12 Oct), and today's closeout is the first with a cause. T-52260, Brook Row: 3.
Accept the 29 clean jobs
Accepted 29 closeouts: 21 electricity, 8 water. Each passed all six checks, and its summary, codes, hours and parts are now in the asset history. One of them, pole AS-1940, carries a proposed follow-up for you to approve.
13 screens from the working solution, on its sample data. Pick one to see it large.
What every job type needs before it is accepted, the tolerances, the code standard and who decides are rules the depot sets. Every change is a new rule version.
| Setting | Default | Choose from |
|---|---|---|
| Required at closeout, by job type | Corrective · LV fuse operation: failure mode, cause, activity, before and after photos, phase voltages after restore, timesheet match, parts vs van stock | Set per job type — water main repair needs pressure after repair; meter exchange needs old and new meter reads; pump fault needs running current |
| Labour against plan | ± 25 % | ± 15 % · ± 25 % · ± 40 % — over this, the job also goes to the supervisor |
| Materials against plan | $50 or 20 % | $25 or 10 % · $50 or 20 % · $100 or 30 % — value or share, whichever is larger |
| Part seen, not booked | Propose to the crew | Propose to the crew · Hold the job until confirmed |
| Photo reader confidence | 0.85 | 0.75 · 0.85 · 0.95 — below this, the photo reader asks rather than proposes |
| Code standard | Lumina failure codes v4 | Lumina failure codes v4 · v3 (retired) — mapped to ISO 14224 failure mode, cause and activity |
| Remind the crew | After 4 hours | After 2 hours · after 4 hours · next morning — if a send-back is not answered |
| Accept clean jobs automatically | Off | Off — the planner accepts them in one click · On above a score of 95 |
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 arithmetic or estimate: $262K = 52,000 jobs a year × 5.8 % with a part seen but not booked × $87 average, the fictional Lumina pilot's inputs; the planner's-morning split is our estimate · “typical” = published figures: truck roll $250–$500 per dispatch, a conservative estimate (S&C Electric, cited by Ventavid). Context: IFS Closeout and Improvement Manager; Siemens Asset Essentials work order auto-summarization; ISO 14224 failure codes. Lumina Utilities and the people, crews, assets and work orders named on this page are fictional — characters and sample data in the working solution.
It is the check a planner or supervisor does on a completed work order before it goes into the asset's history: are the failure mode, cause and activity coded, are the hours and parts booked, are the photos and readings there, and has every defect the crew saw been raised. Work Order Closeout Validation does that check on every closed job with six agents, and the planner accepts or sends back.
The moment a crew closes a job, the agents read its codes, notes, photos, readings, timesheets and van-stock issues against the rules for that job type. Each job becomes a row of six marks with a quality score. Clean jobs are accepted in one click; for the rest, the planner sees each gap with its evidence and sends it back to the crew, who confirm from the van.
Yes. The photo reader reads part labels in the crew's photos and the reconciler matches van-stock issues to the job. In the sample case a new LV fuse-link 100 A gG was legible in photo 2 and issued from van V-114 at 03:58, but not on the parts list. The line is proposed for the crew to confirm; the agent never books it.
Codes are suggested from your own code standard — in the working solution, Lumina failure codes v4, mapped to ISO 14224 failure mode, cause and activity. The notes reader cites the sentence behind every code, says so when the notes do not support a cause, and never writes a code to history without a person accepting it.
The follow-up finder reads the notes and photos for defects, checks for open work on the asset and drafts the work order with a priority and an estimate. In one night it found six, from a cracked crossarm to a boundary box lid that was a trip risk. Nothing is raised until the planner approves it.
Yes. The working solution checks electricity jobs — LV and HV faults, cable faults, inspections — and water jobs — main repairs, valves and hydrants, meter exchanges, pump faults — each with its own required codes, photos and readings, such as phase voltages after restore or pressure after a main repair.
Yes. Agents propose and never book parts, change timesheets, close jobs or raise work on their own. The planner accepts, sends back and approves follow-up work; the crew lead confirms anything an agent proposed; labour over tolerance goes to the supervisor and repeat failures to the reliability engineer.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your closeout procedure, failure code standard, job types and a sample of completed work orders — and it goes live once every quality gate has passed. We will walk you through it on your own closed jobs first.
From paper redline to GIS edit in days, not months, every edit traced to the pen.
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 Work Order Closeout Validation on a sample of your own completed work orders, photos and timesheets.