Guide
How to verify smart meter installations remotely
You verify smart meter installations remotely by requiring evidence with every job (scanned serials, seal numbers, before and after photos, a GPS fix and a timestamp) and having a supervisor review each one daily. Automatic checks catch duplicates and bad locations, and a small number of physical spot checks keeps everyone honest.
The problem with "installed"
On most rollouts, "installed" means someone said so. A count in a sheet, a photo in a WhatsApp group, a crew leader's word. That works until the HES can't find a meter, a consumer says nobody came, or an agency's invoice doesn't match what the DISCOM can see.
Sending someone to check every installation isn't an option at rollout scale. So the job has to arrive with enough evidence that a supervisor can verify it from a desk, and a few physical checks are kept for what the evidence can't prove.
Step 1: Decide what evidence each job must carry
Agree this with the DISCOM or AMISP before work starts, and make every item required in the app.
| Evidence | What it proves |
|---|---|
| New meter serial, scanned from the barcode | The right meter, with no typing errors |
| Old meter serial and final reading, with a photo | What was removed, and the reading billing closes on |
| Seal numbers | The meter was sealed, and which seals were used |
| Photos of the installed meter, showing serial and seals | The meter is on the wall, as recorded |
| A wider photo of the installation | Where the meter is mounted |
| GPS fix with accuracy | The job was done at that premises |
| Timestamp and technician | When, and who |
Add steps if your contract needs them, such as a consumer confirmation or a DT code. Just make sure each one is required, not optional.
Step 2: Let the system do the boring checks
Some checks don't need a person at all:
- Duplicate serials. A meter serial already recorded elsewhere in the project should be refused at submission.
- Location. If the service point already has a location from billing or indexing, compare the job's GPS with it and flag jobs that are too far away.
- GPS accuracy. Flag fixes with poor accuracy for a closer look.
- Completeness. Required fields and photos mean an incomplete job can't be submitted in the first place.
Step 3: Review every job, every day
A supervisor opens each submitted job and checks it against a short list: serials readable in the photo and matching the scanned value, seals visible, readings plausible, location sensible. Then approve it, or reject it with a note specific enough to act on, such as "seal photo blurred, retake on next visit". Never just "wrong".
Daily matters. A rejection the next morning means the crew is still in the area. A rejection three weeks later means a special trip.
Step 4: Watch the patterns, not just the jobs
Individual jobs tell you about one installation. Patterns tell you about a crew or an agency:
- Rejection rate by crew, team and agency
- The most common rejection reasons
- Jobs submitted far from their service point
- Crews whose jobs are consistently completed at odd hours or impossibly fast
A crew with a high rejection rate needs training. A crew with a zero rejection rate and suspicious timings may need a visit.
Step 5: Keep a few physical spot checks
Remote verification shrinks site visits; it doesn't remove them. Pick a small random sample of approved jobs each week, plus anything the patterns flagged, and send someone to look. Tell crews this happens. The point is less about catching people than about making the evidence trustworthy.
Step 6: Pay on verified work
If agencies or installers are paid per meter, pay on approved jobs, not reported ones. That single rule does more for evidence quality than any amount of training.
Doing this in PIEBENCH
PIEBENCH captures the evidence in the Scout Android app, offline, with serials scanned from the barcode and duplicates refused within a project. Supervisors review jobs in an approval queue and reject with notes the crew sees in the app. On the Pro plan ($499 a month), auto-approval rules can compare a job's GPS with the service point's location from billing, among other checks. Smart Views give approved, pending and rejected counts by crew, team and contractor. See installation verification.
Built to integrate with your systems
HES, MDM, SAP IS-U, Oracle CC&B and billing systems, over REST, OAuth or a direct SQL connection. How integrations work
Questions buyers ask
How can I verify smart meter installations without visiting the site?
Require evidence with every job (scanned serials, seal numbers, before and after photos, a GPS fix and a timestamp), have a supervisor review each job daily, and use automatic checks for duplicate serials and locations. Keep a small random sample of physical spot checks.
How do I stop fake meter installations?
Make the evidence hard to fake and check it: serials scanned from the barcode, duplicate serials refused, GPS compared with the premises location, timestamped photos showing serial and seals, and supervisor approval before a job counts. Paying only on approved jobs removes most of the incentive.
What GPS accuracy is good enough for meter installation proof?
It depends on how close premises are to each other. In dense urban lanes, a fix that is off by tens of metres can point at the neighbour's house. Record the accuracy with every fix, and flag fixes that are too coarse to tell premises apart.
Should installers be paid on reported or verified installations?
On verified ones. Paying on approved jobs means evidence quality is the installer's concern as well as the supervisor's, and it removes the incentive to report work that was not done or was done badly.
How many physical spot checks are needed?
There is no single right number. Start with a small random sample of approved jobs each week plus every job your patterns flag, and adjust. If spot checks keep agreeing with the evidence, you can sample less.
Free trial · no card
See it with your own crew
Start a free trial, or ask us for a walkthrough of your rollout.