The short version for the service desk.
- The blind spot: OCPP telemetry is the charger's own account of itself. It cannot see the connector, the cable, the screen, the RCD, physical damage or the car's dashboard.
- The second sensor: the driver or site contact already standing at the unit, recording six things in a set order, guided, in about two minutes.
- The payoff: session, app, tripped-protection and vehicle-side faults close without a visit. The visits that remain leave with the right module on the van.
- Who this is for: service managers at charge point operators and support leads at installers, from home units to rapid hubs.
The dashboard says Available. The driver says the screen is dark, the cable will not lock, or "Preparing" has been on the display for six minutes. Both of them are telling the truth.
That gap has a number now. ChargerHelp's 2025 reliability report, built on more than 100,000 sessions across 2,400 chargers, found a first-time charge success rate of 71% against reported uptime of 98.7 to 99.9%, and more than one third of the failures happened on chargers that appeared operational, according to the report summary on Automotive World.
I've listened in on helpline calls where the agent had the OCPP status open and still could not tell whether the fault was a pressed emergency stop or a cut cable. The information that decided the ticket was standing four metres from the driver's phone, and nobody asked for it.
This post covers what backend telemetry does not show, what to ask the person at the charger to record and in which order, which faults close without an engineer, and how the visit that still happens gets the right part on the van.
If you run service for a charging network or the support desk at an installer, you know the engineer who drove 90 minutes to reset an RCD, and the "Available" unit that had no cable for two days. Venta Capture, a product of VentaVid, was built for that gap: the driver or site contact records what your second-line team needs to see, step by step, no app, and it lands as a structured case against the unit ID. Start for free and build the flow around the fault your desk logs most.
The fault class telemetry cannot see
Telemetry is good at what it was built for. Offline units, firmware versions, backend authorisation, meter values, a controller error code, the session log. If a whole site drops at once, the logs point at power or networking before anyone picks up a phone.
The remote diagnostics guide for field service covers the general version of that logic.
The faults that eat engineer days are the ones the controller reports as fine, because the controller is fine. The problem is everywhere around it.
Six things the backend does not see:
- The connector face and latch: bent pins, debris in the socket, a latch that no longer holds, a driver trying a CCS lead on a CHAdeMO car.
- The cable: cut, crushed by a kerb, run over, kinked at the strain relief, or simply gone. A tethered lead can be missing for days on a unit reporting Available.
- What the screen shows right now: a frozen interface, a dark display with the LED still alive, a message in the wrong language, an "out of order" sticker somebody stuck on last week.
- The supply side: a tripped RCD or RCBO, an isolator switched off by a cleaner, an emergency stop button pressed by a child.
- Physical damage and water: a snapped holster, a cracked screen, condensation behind the glass, a bollard knocked into the cabinet.
- The car's own display: scheduled charging set in the vehicle, a charge limit already reached, a 12V battery warning, a vehicle-side fault that has nothing to do with your unit.
The numbers say this is the majority case. In the J.D. Power 2025 EVX Public Charging Study, 14% of owners visited a charger without managing to charge, and "the charger being out of service or not working properly" was the reason in 60% of those failed visits, per the study summary.
And the cheap fix is often the whole fix. Franklin Grid puts it bluntly: communication failures inside the charger leave the operator unable to detect or react remotely, 50 to 70% of the time a power cycle is all that is needed, and rolling a truck to flip a breaker costs $500 to $700, by their estimate.
So the question for remote triage is not "what does the unit say about itself". It is "what does the unit look like right now, to the person standing at it".
What the driver or site contact records, in order
Asking a driver to "send some photos" gets you a blurry shot of the bay from the driver's seat. The order and the instructions have to come from your side, which is the whole idea behind guided capture: your second-line engineer's checklist becomes the steps on the participant's phone.
Each step exists because it settles one decision the desk would otherwise guess at. The table is the flow as the participant sees it, top to bottom.
A few notes on running it:
- Two yes/no questions route the flow. "Did the app or card authorise?" and "Did the cable lock?" split a session problem from a hardware problem before the first photo, and the desk reads those two answers before anything else.
- Step 5 is photograph only. Nobody opens an enclosure. At a home or workplace install, the customer's own consumer unit is in scope; at a public site it is the host's isolator, per the site rules you set in the flow.
- The spoken description matters as much as the picture. "It clicked twice and then went dark" is diagnostic. Venta Capture transcribes the audio, so the desk can search it later.
- The unit ID is the case reference. Read it from the label or scan the QR in step 6, and the submission files itself against the right asset in your ticket.
The whole thing takes a driver about two minutes on a motorway forecourt, and a site host less, because they know where the isolator is. The guided photo capture post goes through how to write the per-step instructions so an untrained person gets them right the first time.
How the link reaches them depends on the channel you already run. In the UK, the 24/7 staffed helpline required by the Public Charge Point Regulations 2023 is the obvious sender: the agent texts a secure capture link while the driver is still at the unit. For installers, it goes out from the support inbox by SMS or WhatsApp.
For a site host, a QR code on the unit or on the site signage does the same job without anyone on the phone.
What comes back is one capture flow submission rather than six loose files in an inbox: the answers to the routing questions, every photo and clip in step order, the transcript, the receipt time and device context, all under the unit ID. That is the format a second-line engineer can review in ninety seconds, at 07:30, for the fifteen tickets that came in overnight.
Faults that close without a visit
Evolt's service team, which completed more than 10,000 on-site visits in 2025, reports that its second-line diagnostics team resolves over 40% of faults remotely, according to its managing director on EV Infrastructure News. That is telemetry plus experienced people. Add what the driver can show you and the list of faults that never become a truck roll gets longer.
With the six captures in hand, four fault families close from the desk. Each one below names the evidence that closes it.
Session and authorisation faults. The app screenshot and the screen clip show a card not registered, a payment declined, a driver on the wrong network's app, or a tariff that needs accepting. The desk fixes the account or talks the driver through it, and the case closes with the evidence attached. Squote's UK fault-finding guide notes that most customer-reported "faults" turn out to be scheduling or Mode 3 protocol mismatches rather than hardware, in its breakdown of fault categories.
Tripped protection and pressed stops. The step 5 photo shows the RCBO down or the e-stop in. Where the participant is allowed to reset it (their own consumer unit, or a host's isolator under the site rules), the desk sends the reset instruction and asks for one more photo. If it trips again with no car connected, that is a supply fault and a real dispatch, now with the reason on the ticket instead of "customer says charger dead".
Vehicle-side refusals. Step 4 catches the car that has scheduled charging set for 02:00, the battery already at its configured limit, or a vehicle fault light. Your unit is fine. The driver gets told, the case is logged as vehicle-side, and the reliability record does not take a hit it did not earn.
Frozen interfaces on a live controller. A screen that is alive but stuck, with the backend showing the unit online, is the case where a remote reboot works. Without the clip you are guessing whether there is anything to reboot into. With it, you push the restart and ask the driver to try again while they are still there.
Venta Capture reports that 23% of customer problems are solved without dispatching a technician. In a network where a wasted visit costs a few hundred pounds or dollars before the engineer's day is counted, that share is the budget line.
The reason this works at 2 a.m. on a service area is that nobody on your side has to be on a call. The driver records now; the second-line engineer reviews when they start.
The customer support use case explains that asynchronous loop for support desks generally, and the remote troubleshooting glossary entry gives the short definition.
Want to see it on your own fault categories? Book a demo and bring last month's no-fault-found list.
When the van still rolls, put the right module on it
Some captures end the debate the other way: the cable is on the ground in two pieces, the connector has a pin folded flat, the cabinet has a bollard in it. The visit is happening. What the evidence changes is who goes and what they carry.
The cost of a wrong first visit compounds, because the second visit inherits the first visit's delay. Aquant's 2025 field service benchmark found that a failed first visit adds two more visits on average and stretches resolution by 14 days, as reported by 24x7 Magazine. Evolt's 90%+ first-time fix rate is the number that separates a service operation from a taxi service for engineers.
Two practical points for the planner:
- The label photo is the parts list. Model and serial from step 6 tell you which generation of comms board or which cable length before anyone opens the van stock system. That is the same "know before you go" idea behind the pre-visit assessment post, applied to a charger.
- A missing detail costs one retake, not one visit. If the first submission has the cable but not the label, the desk requests a retake in one click, and the participant gets a link that asks only for what is missing.
The work order then carries the view link. The engineer opens it in the van, sees the same six captures the desk saw, and arrives already knowing which side of the unit to open.
The record: helpline log, reliability report, warranty claim
For UK operators the evidence has a second life. The Public Charge Point Regulations 2023 require rapid charge points of 50kW and above to hit 99% reliability as an annual average across the network, with the first report covering 2025 due by 14 January 2026, as Safer Charging summarises. The same page cites a 2025 Monta survey of more than 200 UK operator decision-makers in which only 3.9% said they currently meet the threshold.
The regulations also require a staffed 24/7 helpline and record-keeping on every contact, with fines for non-compliance running from £10,000 to £250,000, according to the RAC.
A sealed capture attached to the helpline contact turns "driver reported a fault at 23:40" into a timestamped record of what the unit looked like at 23:40. That record does four jobs:
- Hardware warranty. A clip of the frozen screen with the firmware version visible and the receipt time logged is what the manufacturer's warranty desk asks for anyway.
- Vandalism and insurance. DeterTech incident data shows EV charger theft and vandalism more than doubled in 2024 versus 2023, per Formula Space's summary. A cut cable photographed by the first driver to find it, with time and location context, is the start of the claim.
- Site host disputes. When the isolator was off because the host's cleaner switched it, the photo settles who owns the downtime.
- The reliability report. Vehicle-side and site-caused failures documented at the time are easier to classify correctly when the annual numbers get compiled.
Be precise about what the seal proves. The receipt timestamp proves when your system received the submission, not when the cable was cut. Device context and integrity signals are reasons for a human to look closer, never a verdict on their own.
The evidence integrity entry explains the control points, and the proof of work post shows the same mechanism used at the other end of the job, when the engineer documents the repair.
Ready to try it on one site? Start for free, build the six-step flow, and send the link to yourself first.
Setting it up for a charging network or an installer desk
The pieces you need already exist in Venta Capture's product; the work is deciding your steps and your routing.
- Start from a standard flow and adjust it. Your team edits the steps, the example photos and the per-step instructions in the builder. Rename "connector" to whatever your engineers call it.
- Use the unit ID as the case reference. Scan it in step 6 or pre-fill it when the helpline generates the link, so every submission files against the asset.
- Route by site or region with an SLA in hours. Notifications go to the second-line queue with the view link, or to a named engineer's phone, on new submission, on claim, on status change. The escalation path and SLA entries cover the vocabulary.
- Branching and integrations are set up around your process. A flow that routes differently for AC versus DC units, or that pushes the case into your CPMS or ticketing tool, is a custom setup VentaVid scopes with you on a call. The free plan gets you the standard flow live the same day.
What not to expect: Venta Capture does not diagnose the charger, does not speak OCPP, and does not replace your CPMS or monitoring. It puts a second, asynchronous source of evidence next to the telemetry, and it makes sure that source arrives in the order your engineers need.
If you are weighing it against live video assistance tools, the remote inspection software overview sets out where each fits; the short answer is that a live session needs an agent at the same moment, and a helpline at 2 a.m. rarely has one to spare.
Other distributed assets with the same problem: solar inverter faults and wind and utility-scale solar. The general case is in remote visual inspection, and how this sits next to your FSM system in remote inspection software vs field service apps.
Frequently asked questions
Can a driver diagnose an EV charger fault remotely?
No, and they should not be asked to. The driver records what the unit, the cable and the car look like right now; your second-line engineer makes the call. The driver's job is two minutes of guided capture, not troubleshooting.
What should I ask a driver to photograph when a charger will not start a session?
In order: the connector face, the full cable, the screen with the message read aloud, the car's own charge screen, the RCD or e-stop if it is safely visible, and the unit label with the surroundings. Two questions first: did the app or card authorise, and did the cable lock.
Which EV charger faults can be fixed without sending an engineer?
Account and authorisation problems, tripped protection that can be safely reset by the person on site, vehicle-side refusals such as scheduled charging, and frozen interfaces on a unit the backend can still reboot. Evolt reports over 40% of faults resolved remotely with telemetry and a second-line team alone.
Why does the backend show Available when the charger is not working?
Because the controller is reporting its own state, and the controller may be fine. A cut cable, a broken latch, a dark display or a pressed e-stop can all leave the unit online and unusable. ChargerHelp found more than a third of failed sessions on chargers that appeared operational.
Should the driver reset the RCD or breaker themselves?
Only where they are entitled to: their own consumer unit at a home install, or a host's isolator under rules you set per site. Nobody opens an enclosure. If the device trips again with no car connected, that is a supply fault for an electrician, not a reset.
Does this replace OCPP monitoring or the CPMS?
No. Telemetry stays the first line for offline units, firmware and backend faults. Guided capture adds the physical layer telemetry cannot see, and the case can be pushed into your ticketing or CPMS through the API or a custom integration.
Do drivers need an app to send video?
No. The link opens in the phone's browser, with no download and no account. The flow runs in the driver's own language, and a desktop visitor gets a QR code to continue on their phone.
Talk to us
If your engineers keep arriving at chargers that turn out to be fine, or at chargers that needed a part they didn't bring, the fastest way to see the difference is to run one live fault through a capture link.
What it takes to start.
- Free plan, no credit card, and the standard flow is live in about 10 minutes.
- Stuck on the steps? Book a free setup call and we build your first flow together.
- Custom routing or a CPMS integration is set up around your process, scoped on a call.
The VentaVid team builds Venta Capture for service desks that need to see the site before they send someone.
Book a demo and bring a real ticket. We will map the six steps to your fault categories on the call.