What Trigo's own documentation says its retail computer vision does not cover, written for loss prevention and store systems teams running due diligence. Every constraint carries its documentation, its operational impact, and its workaround.
Every constraint below comes from Trigo's own current documentation, checked on August 17, 2026. Where Trigo documents nothing, this page says so rather than guessing, because an absence in a document is a question to ask rather than a finding to claim. Spot AI sells a competing platform, so nothing here rests on an anonymous source or an aggregated user rating, and this page corrects something Spot AI has itself published about Trigo. We wrote that its camera requirement was not publicly specified and that, because its origin is fully instrumented checkout-free stores, a buyer should establish early what camera coverage a deployment assumes. For the loss prevention product that is wrong: Trigo's own wording is that it works with any camera system regardless of brand or specifications, using the existing camera network with no additional camera investment required. The caution holds for the autonomous retail product only, and constraint 04 below draws that line.
The loss prevention product runs on the cameras already in the ceiling. The autonomous product does not, and Trigo publishes the difference. Both stop at the same place: the checkout, and a notification.
A constraint list is only useful next to an honest account of the product.
It works with any camera system regardless of brand or specifications, using the existing camera network with no additional camera investment required. Most loss prevention deployments are published as needing minimal new hardware and completing in as little as two weeks, on a single in-store workstation. For a chain with 400 stores and 20 years of mixed hardware, that removes the largest line from a rollout budget before the conversation starts.
Trigo integrates with the store's existing point of sale and self-checkout to follow items from aisle to checkout to exit, so a non-scan arrives as a reconciled event with the item and the lane attached rather than as a clip somebody has to interpret. The named detections are legible and short: hidden products, self-checkout non-scan, checkout skipping, manned checkout monitoring and Scan and Go losses.
No biometric identification, automatic face blurring, fully anonymized shopper tracking and compliance stated against GDPR, described by Trigo as privacy by design. Behind that sit over 60 million shopping activities processed annually, up to 100,000 products identified and shoppers monitored across thousands of cameras, with Tesco in the UK and REWE in Germany named among the retailers running its autonomous stores.
Each one is a consequence of how the platform is designed rather than a defect. What matters is whether it collides with your starting point.
The named detections are hidden products, self-checkout non-scan, checkout skipping, manned checkout monitoring and Scan and Go losses, with mis-scans, fake scans, payment skippers and walkouts named on the self-checkout page. No safety, operations or perimeter detection appears on any published page, so the stockroom door, the loading bay, the yard and after-hours entry are outside the published scope.
A grocery chain whose exception report says self-checkout is buying exactly the right product. The same chain's incident log usually also carries back-door shrink, a delivery bay dispute and an overnight entry, and none of those sit on the till. Covering them means a second platform reading the same cameras, with a second integration and a second renewal date.
Put the exception report and the incident log side by side before the shortlist closes, and mark which lines are till-adjacent. Keep Trigo for the till-adjacent ones, where it is precise, and price the rest as its own decision rather than assuming the checkout product will grow into it.
Trigo publishes real-time notifications to staff and security carrying video evidence, plus the option of a prompt on the self-checkout screen showing the shopper the image of the item. No speaker, talk-down, strobe or horn behavior is publicly specified. The documentation covers what the store is told, not what the store does automatically.
The screen prompt is a genuine intervention at the lane and it works without anybody moving, which is the strongest part of this design. Away from the self-checkout, and on a thin shift or after closing, the sequence ends with a notification, so the value of a detection is capped by who can reach the aisle in the time it takes to leave the store.
Measure how long it actually takes a colleague to reach the lane on your thinnest shift and use that number, not the alert latency, to size the benefit. Lean on the screen prompt where it applies, and for closed hours price a platform that drives the speakers already installed in the store rather than assuming a notification covers it.
Trigo names integration with the store's existing point of sale and self-checkout as part of the requirement for the loss prevention product rather than as an optional extra, alongside a connection to the camera network already in store and a single in-store workstation. Which point-of-sale platforms and versions are already integrated is not publicly specified.
The reconciliation only works if the transaction data arrives cleanly, which means the till estate owner belongs in the room from the first scoping call rather than the third. On a chain running two point-of-sale generations across acquired banners, that integration is the item most likely to move the go-live date, and it sits with a team that did not ask for this project.
Name your point-of-sale platform and version in the first conversation and ask for a reference customer already running it. Give the integration an owner, a date and a test store inside the pilot, and confirm what the detections do on a lane where the transaction feed is delayed or missing.
The loss prevention product states it works with any camera system regardless of brand or specifications. The autonomous retail product is published with different conditions: ceiling-mounted cameras, compact edge computing units and a reliable internet connection, in stores stated at up to 1,200 square meters. Trigo is explicit about the difference rather than blurring it.
Which product a quote is actually for decides whether there is hardware in the number. A frictionless-checkout pilot carries a ceiling camera installation and an edge unit per store, which is a capital project per site, while a loss prevention rollout on the same estate can run on the cameras already in the ceiling. Reading the two as one platform is how a budget doubles.
Make the quote name the product. If it is autonomous retail, price the ceiling camera work and the edge units per store and check your store footprints against the published 1,200 square meter figure before scoping. If it is loss prevention, hold the vendor to the no additional camera investment wording.
GDPR compliance is stated and the privacy design is documented in detail. No SOC 2, ISO 27001 or equivalent certification statement appears beside it. Where full-resolution video lives is not publicly specified beyond a single in-store workstation and what Trigo describes as optimized on-prem clustering, and no retention period or export path is published.
The privacy answer is strong enough to clear a works council, and it is not the answer a security review asks for. A retail IT team needs the certification pack, the retention schedule and the export path for evidence, and none of the three can be read off a public page, so they surface late in the cycle when they are most expensive.
Request the certification pack, the retention schedule and the evidence export path in one email at the start of the evaluation rather than the end. Ask specifically what the in-store workstation holds, for how long, and what happens to it if the store loses connectivity or the contract ends.
The same five constraints in one view, sized to paste into an evaluation document.
Swipe the table sideways to see every column.
Trigo data comes from Trigo's own loss prevention, self-checkout, autonomous retail and privacy documentation, checked on August 17, 2026. Gaps are marked as not publicly specified.
If the exception report says self-checkout, if non-scans and walkouts are the pattern, and if the point-of-sale data is already clean enough to reconcile against, most of the constraints above stop applying. Trigo is a precise answer to a precise problem, it runs on the cameras already in the ceiling, and the published two-week deployment and privacy-by-design position make it one of the easier products in this market to get past a works council and into a rollout plan.
Spot AI covers the store rather than the lane, which is why the two constraint lists barely overlap. 15+ pre-trained Video AI Agents run across the whole estate: cash register theft, after-hours intrusion, vehicle break-in, fire, personal protective equipment, forklift near-miss, falls and crowding in hazard zones, so the back door, the loading bay and the yard are inside the same deployment as the sales floor. Iris builds anything else in natural conversation in about eight minutes.
The response and the architecture close the other two gaps directly. An event at 02:00 is answered on site through talk down, strobes and horns on standard speakers rather than by a notification waiting for somebody to read it. Any ONVIF or RTSP IP camera works at full functionality and legacy analog comes in through the Intelligent Video Recorder, and full-resolution video stays on the IVR in the store with only event metadata crossing the network, which keeps the retention and residency answers inside the building.
The useful question is not which platform has fewer constraints, but whether your loss is concentrated at the lane or spread across the building.
None of that makes Spot AI the better answer at the checkout, and on a self-checkout shortlist it is not the same product. Trigo reconciles what a shopper picked up against what was scanned by integrating with the point of sale, and Spot AI does not perform transaction-level reconciliation, so a Spot AI detection at a lane arrives without the item and the basket attached. Trigo also publishes a two-week deployment figure and a store-size limit for its autonomous product, where Spot AI publishes no deployment time at all, and frictionless checkout has no Spot AI equivalent whatsoever. A constraint list earns its keep by lining each platform's shape up against where your loss actually happens.
A live pilot on your store cameras answers in a week what a spec sheet cannot.
Customer-reported outcomes from named Spot AI customers.
All Star Elite cut cash shrink from about 6% to 1% across its stores and reported investigations more than 50% faster.
GO Carwash reported a 54% increase in membership conversion at pay stations after closing the service gaps the video surfaced.
Blackmon Oil Co. runs a vehicle loitering filter across its round-the-clock stores so the lot stays watched with a smaller overnight crew.
"I like the fact that you can build a case to share with insurance companies. It's easy to navigate and with license plate recognition, incidents are much easier to find."
Five, all documented. The named detections stop at the checkout, so safety, operations and perimeter events are outside the published set. The documented response is a notification with video evidence plus an optional self-checkout screen prompt, with no speaker, talk-down or strobe behavior published. Point-of-sale integration is named as part of the requirement rather than as a setting. The camera-reuse claim covers loss prevention and not autonomous retail. And the certification pack, the retention period and the export path are all request-only.
Yes, and for loss prevention the claim is the broadest in this category, in Trigo's own wording: it works with any camera system regardless of brand or specifications, using the existing camera network with no additional camera investment required, with most deployments completing in as little as two weeks on a single in-store workstation. The autonomous retail product is the exception and Trigo says so, calling for ceiling-mounted cameras, compact edge computing units and a reliable internet connection.
No, and it publishes the position rather than leaving it to a policy document: no biometric identification, automatic face blurring, fully anonymized shopper tracking and compliance stated against GDPR, described as privacy by design. That is usually enough to get a works council or a data protection officer comfortable early, which is not true of every product a retailer might put on the same cameras.
Not in the published set. Hidden products, self-checkout non-scan, checkout skipping, manned checkout monitoring and Scan and Go losses are what the pages name, and no perimeter, safety or operations detection appears anywhere. So the loading bay dispute, the back-door shrink and the overnight entry sit outside scope, and covering them means either a second platform on the same cameras or a platform whose set spans both from the start.
It depends on where the loss actually sits. If the exception report points at the lane, Trigo reconciles picked against scanned better than a general video platform can, and that line is worth keeping. If the loss is spread across the building and the overnight hours, a platform such as Spot AI fits: 15+ pre-trained Video AI Agents run on any ONVIF or RTSP camera plus legacy analog through the Intelligent Video Recorder, footage stays in the store, and an event is answered with talk down, strobes and horns.