The short version for the service desk.
- The move: before anyone books travel, the operator films the machine for you, guided step by step from a link, and the desk reviews it when it lands.
- What to ask for: five things, in order. The machine running or failing, the HMI or panel message, the type plate, the affected assembly, and the sound of a start attempt.
- The number: avoidable dispatch runs at 24% for the bottom fifth of service organisations and 3% for the best, according to Aquant's 2025 benchmark. Most of that gap is information.
- Who this is for: service managers at equipment OEMs and machine dealers, and the maintenance contractors who look after fixed production equipment for them.
"Line two is down. There's an alarm on the screen. It was fine yesterday and nobody touched anything."
That is the whole ticket, and if you run a service department at an equipment manufacturer or a machine dealer you get a version of it every day. No alarm number and no serial. No idea whether it's the drive, the sensor, the guard switch or the operator, and the machine is four hours away.
So the desk does what desks do. It rings back, asks the questions the caller can't answer, and eventually sends an engineer with a van full of guesses. Sometimes the engineer presses reset and drives home. Sometimes the fault is real and the part on the van is the wrong revision.
Industrial equipment service video is the fix for the first half of that problem: getting the desk eyes on the machine before it commits an engineer, and getting the engineer's own eyes into the record once the job is done.
If you manage a service desk for fixed production equipment and your hotline still dispatches on a sentence, this is the workflow we built Venta Capture, a product of VentaVid, to run: the operator records what your second line needs to see, guided step by step, and it arrives as a structured case tied to the serial. Start for free and send a capture link to your own phone before you read on. Nothing to install on either end.
In this post:
Why a description is the worst input a service desk can work from
A fixed machine is the easiest thing in field service to see. It doesn't move, the alarm is on a screen bolted to it, and the type plate is riveted to the frame. Even the noise comes from one place.
And yet what reaches your desk is a paraphrase. The operator tells the shift lead, the shift lead tells plant maintenance, plant maintenance rings your hotline, and your first line types a summary for second line. Three translations, and the alarm number fell out somewhere around the second one.
The maintenance textbooks know this. ATS's troubleshooting guide puts "talk to the person who requested maintenance support and to whoever first spotted the issue" at step three, then warns about recency bias in what operators remember. What they recall is the last thing that happened. What your engineer needs is the first.
The desk isn't short of expertise. It's short of the two details that decide whether this is a remote fix, a one-hour visit or a two-day one. Which alarm. Which sound.
I've sat next to second-line engineers doing this triage, and the tell is always the same. At some point the caller holds the phone up to the machine and says "can you hear that?" Everything the engineer needs is right there, and the only channel is a phone speaker and someone else's vocabulary.
What it costs when the guess is wrong is well documented, and it's the customer's number before it's yours. Siemens' 2024 downtime study puts the average large plant at 27 hours of unplanned downtime a month across 25 incidents, with an automotive line costing up to $2.3 million an hour when it stops. Your service contract lives or dies on how fast you turn those hours around.
On your side of the contract, Aquant's 2025 benchmark of 160 service organisations found avoidable dispatch at 3% for top performers and 24% for the bottom fifth, and a failed first visit adding two more visits and 14 days to resolution. One in seven truck rolls is unnecessary.
The remote-access answer to this is real, and if your machines ship with a gateway, use it. HMS Networks, who sell those gateways, reckon 60 to 70% of cases need no on-site visit once an engineer can see the HMI and the PLC remotely, and cost a single engineer mission at roughly EUR 3,525 all in.
But most of the installed base has no gateway. It's ten years old, it's behind the customer's firewall, or it's a mechanical fault a PLC can't see. The phone in the operator's pocket has none of those limits, which is the whole argument for remote diagnostics without sensors. This post is about what to do with that phone on a production machine specifically.
The five things to ask the operator to film
"Send us a video" already happens over WhatsApp at most OEM desks, and it mostly produces 40 seconds of a blurry cabinet with the alarm just out of frame. The difference between that and a usable case is that your desk decides what gets filmed, in what order, with an instruction on screen for each step.
For fixed production equipment, five items cover nearly every dispatch decision, and the order matters because the operator's patience runs out around item four. Put the plate first if your parts desk is the bottleneck; otherwise this order.
- The machine running, or failing to. Twenty seconds of the cycle up to the point it stops. Jams, misfeeds, a servo that hunts, a conveyor that stalls under load: the engineer sees the behaviour instead of hearing an adjective for it.
- The HMI or control panel message. The alarm screen, the alarm list if there is one, and the LEDs on the drive or safety relay. This is the item that falls out of every phone call, and it's the one that most often decides "remote fix" versus "visit".
- The type plate. Model, serial, year, ratings. A scan step reads the barcode or QR sticker with a frame on screen and an example photo of where the plate sits on that model, and falls back to typing if the plate is worn. Everything downstream, from parts to warranty, hangs on the serial number.
- The affected assembly. A close-up of where the operator thinks the problem is: the leak, the belt, the coupling, the sensor with the loose connector. Only with guards that can be opened safely with the machine isolated; the flow says so on screen.
- The sound. A start attempt with the phone held near the assembly, while the operator describes out loud what they hear and where. Bearings, belts, cavitation and chatter are diagnostic to an engineer who has heard a thousand of them, and no form field captures a knock.
Two questions ride alongside the footage, answered by tapping, not typing an essay: what was the machine running (product, material, speed), and what changed (new batch, new operator, maintenance yesterday, a power cut). A yes to "was any work done on the machine in the last 48 hours?" steers to a follow-up; a no skips it.
That list is a guided capture flow, and it takes an operator about three minutes to walk through. Build it once per machine family in the flow builder and every hotline call for that family starts with the same evidence. The guided photo capture post covers why example photos and per-step instructions are what turn an untrained participant into a usable pair of eyes; the short version is that the expertise lives in the flow, so the operator doesn't need any.
Direction one: the operator captures before you dispatch
The desk-side flow is short, because most of it is waiting.
- The hotline opens the case and sends the link. A secure, personal capture link goes to the operator or the site maintenance lead by SMS, email or WhatsApp, or they scan a QR code on the machine's service screen. No app, no account, works in the phone's browser in the operator's own language. That is the whole of requesting video by link.
- The operator films the five items. Instructions on screen per step, an example photo of the type plate location, a counter of required items left. A desktop visitor gets a QR code to continue on the phone, so nobody uploads last week's footage from a laptop.
- The case lands in the inbox. Video and photos, the tapped answers, a transcript of what the operator said out loud, the timestamps and device context, all under your own case reference, with a "your reference" field for the ticket number your system already assigned.
- Second line reviews when they're free. If the plate is unreadable, a retake request goes back in one click, and the retake links to the original and records why. Reviewing, retaking and approving is described on the review and retake page.
- The desk decides. Fix it over the phone with the operator now that both sides are looking at the same alarm. Dispatch with the right engineer and the right parts. Or escalate to the product specialist with a note, without re-explaining anything.
The part that changes the operating model is step four: nobody had to be free at the same time. Live video calls and AR sessions work, and they stay in the toolbox for the cases that need an expert steering the camera. But a live session is a meeting. It needs the operator, the engineer and a working connection at the same moment, which on a three-shift plant with a support desk in another time zone is exactly what you don't have.
Asynchronous capture removes the calendar. The night shift films at 02:00, second line reviews at 08:00 with coffee, and the engineer who does travel at 10:00 already knows it's the encoder and has one on the van. Send now, capture later, review when ready.
That is what a pre-visit assessment looks like on a production machine, and the pre-visit assessment post covers the general mechanics if you want them. Where the desk decides not to travel at all, that's a truck roll saved, and the reduce truck rolls post walks through how to count those honestly. Across support and field-service teams using Venta Capture, 23% of customer problems are solved without dispatching a technician. The rest get an informed visit, which is the larger win on equipment where a wrong part means a second trip.
Want to see what the operator's screen looks like on one of your machine families? Book a demo and bring a real alarm from last week.
Direction two: the engineer captures on site
The second direction gets less attention and pays for the service contract.
Your engineer arrives with the case already open. What they add to it on site is what the customer, your warranty team and the engineer who visits the same serial next year will all want later:
- The failed part next to its own label. Part number, revision, batch code if there is one, damage visible.
- The replacement fitted. Same framing, so the two photos compare.
- The machine running after. Twenty seconds of the cycle that was failing this morning, now not failing.
- The counter or hour meter. Machine hours at the visit, for the warranty window and the maintenance schedule.
- The handover. The customer's maintenance lead confirming on camera that the line is back, plus a declaration checkbox as the last step.
Staff can start a capture themselves in the dashboard without sending anyone a link, so this is a two-minute job at the end of the visit, not admin at the hotel. The before-and-after can sit as two linked cases in one file: the operator's capture from this morning and the engineer's from this afternoon, against the same serial. That's proof of work in its most literal form, and the proof of work post goes into why customers stop disputing invoices when they can watch the repair.
A service engineer's report used to be a paragraph typed in the van; now the report is the footage, and the paragraph is the transcript. The handover step matters most on contract work, where the argument six months later is never about whether you came, but about what state you left it in.
What the record is worth after the visit
Each case carries a complete audit trail: when the link was created and by whom, each time it was opened, when the submission arrived, who claimed it, every note, and the session events behind the recording. On top of that sits a digital seal with a SHA-256 fingerprint and a downloadable manifest, so the file can be verified outside your own system. The sealed submission glossary entry explains the mechanism; the evidence upgrade is the module that provides it.
Where an OEM or dealer service desk cashes that in:
- Warranty claims to component suppliers. Photo of the failed drive next to its label, the alarm history from the operator's capture, the serial and the hours. A warranty claim with a timestamped picture of the failed part beats one with a description, and the supplier's warranty desk is having the same argument with you that you have with your customers.
- Service contracts and SLAs. When the fault was reported, when the desk first looked at it, when it was fixed. Response and resolution against a service level agreement stop being a matter of whose email thread is longer.
- The next fault on the same serial. Six months on, a different operator, the same knock. The case from March is one search away, with the sound in it.
- Disputed invoices. The engineer's afternoon capture shows the line running. The conversation ends there.
One honest caveat that the product itself is explicit about: the server receipt time proves when your system received the submission, not when the fault occurred. Write that into your contract language rather than around it. What the evidence integrity layer gives you is a chain from capture to case that nobody has to take on trust.
Cases can also leave the inbox on their own. A rule can forward a submission to your ticketing or CRM by status, or fire a webhook so the case number, serial and view link land in the ticket your engineers already work from. That's covered on the workflow automation page. Flows with branching between paths and routing into your systems are set up with you around your process; the standard fault-capture flow you can adjust yourself in the builder the same afternoon. The remote inspection software post compares the categories if you're weighing this against a field-service management app.
If the warranty desk is the team that would benefit first, start with a free account, build the engineer's on-site flow, and run it on the next three visits.
What operator video does not do
Some limits, because a service manager who has been sold "remote everything" before deserves them up front.
- It doesn't diagnose. The recording gives your engineer better eyes. The engineer still makes the call, and remote triage is a person's job, not the tool's.
- It doesn't replace PLC access. If a machine has a gateway and the fault is in the logic, log in. Video is for the unconnected base, for anything mechanical or hydraulic, and for the alarm you need to see before you know whether to log in at all.
- It can't see inside a sealed gearbox. Some faults need a visit, and the honest promise is remote-first, not remote-only: fewer blind visits, better-prepared ones.
- It never asks an operator to do something unsafe. No opening guards on a live machine, no defeating interlocks for a shot. The flow's instructions say what to film with the machine isolated, and if the item can't be captured safely, it's skipped and the engineer goes.
- The operator's willingness is finite. Three minutes and five items. A twelve-step flow gets abandoned at step six, which is why the field service flows ship short.
Get those right and the trade is simple: three minutes of the operator's time against a four-hour drive to press reset.
Mobile plant and forklifts are covered separately in construction equipment fault video before dispatch. The identification step is in equipment identification on video.
Frequently asked questions
Does the operator need to install an app?
No. The capture link opens in the phone's browser, in the operator's language, with no account to create. On a computer it shows a QR code to continue on a phone, so the footage is recorded on the device in the session, not pulled from a gallery.
What if the operator films the wrong thing?
The flow tells them what to film at each step, with an example photo, and marks required items. If something is still unusable, second line sends a retake request in one click and the retake links to the original case with the reason recorded.
We already get WhatsApp videos from customers. What changes?
Order, completeness and the record. A guided flow gets the type plate and the alarm screen every time, ties the footage to the serial and your ticket number, transcribes what was said, and keeps an audit trail you can hand to a warranty desk. A WhatsApp clip does none of that.
Does this replace remote access to the PLC or HMI?
No. On connected machines with a logic fault, remote access is faster. Operator video covers the machines without a gateway, mechanical and hydraulic faults, and the first look that tells you which of the two routes to take.
Can the recording be used in a warranty claim?
Yes, with the right framing. The case carries the failed part on camera, the serial, the machine hours, the timestamps and a sealed audit trail. The receipt timestamp proves when the submission arrived, not when the failure happened, so pair it with the alarm history and the operator's account.
How long does it take to set up a fault-capture flow?
The free plan is live in about 10 minutes: pick a standard flow, adjust the steps for your machine family, send a test link to your own phone. If you want a branching flow that routes different fault types to different specialists, that is set up with you around your process on a call.
Is this for the customer's operators or for our own engineers?
Both, and that's the point. The operator captures before dispatch so the desk can decide; the engineer captures on site so the customer, the warranty desk and next year's visit have the record. Same tool, same case, two directions.
Talk to us
If your hotline still dispatches on a sentence, the fastest way to see what changes is to send a capture link to your own phone and film a machine on your shop floor. Five items, three minutes, and you'll know whether your second line could have made the call from it.
What it takes to start.
- Free plan, no credit card. Build a fault-capture flow for one machine family and run it on the next hotline call.
- Nothing for the operator to install. A secure link in their language, recorded on their phone.
- Stuck? Book a free setup call and we build your first standard flow together.
The VentaVid team builds Venta Capture for service desks that would rather see the machine than hear about it.
Start for free, or book a demo and bring last week's worst ticket. The product page has the full feature list if you'd rather read first.