Scheduled, barcode-bound inspection rounds that work offline, flag out-of-range readings at the asset, and trend every measurement point.

You can show that the round happened at the asset, that the out-of-band reading was seen at the asset, and that the number in the trend is the number the technician read.
The inspection programme becomes measurable: which assets keep failing, how long closure really takes, and which defects are still open, instead of a folder of completed checklists.
Upload it as a spreadsheet. Columns are mapped onto form fields, including into repeatable sections where one asset carries several components or test points, and the mapping is stored with the schedule so subsequent uploads follow it. From that point a schedule can generate one prefilled round per asset row on whatever frequency you set. You do not need the register to be perfect first: the mapping is per schedule, so a pilot on one unit does not require the whole estate to be reconciled.
Three mechanisms, used together. A barcode or QR scan binds the record to the asset tag, so the identity is read rather than typed. A Geo Control question stamps the coordinates onto the submission, and a proximity rule compares that against the asset location and can warn, require a justification or block. Where per-field metadata is enabled at dispatch, individual answers carry their own timestamp and geostamp, so a round completed in one burst in the car park looks different in the data from a round walked over ninety minutes.
Yes, and this is the reason repeatable sections matter more than they first appear. Each row of a repeatable section, one condition monitoring location, one test point, one component, is stored individually in the analytics store with its own row index, not squashed into the parent submission. That means a specific point can be plotted across every inspection it has ever appeared in, and derived values like corrosion rate can be computed at the point level rather than the round level.
Yes. An API question calls an external system from inside the form and writes the response into named target fields, including into a specific row of a repeatable section. It can fire when the page loads, when a field such as the scanned asset tag changes, or only when the technician presses the button. Every call is logged with its status, latency and a redacted payload, and the endpoint configuration including credentials is encrypted at rest. The same configuration can be dry-run from the admin side before it is published, so you find out the endpoint is wrong at design time.
Yes. The mobile app is offline-first: forms, reference lists and the compiled rule logic are downloaded to the device before the technician leaves, and submissions including photographs and audio notes are stored locally, encrypted, until connectivity returns. Rules run on the device, so range validation and conditional sections behave identically offline. API questions that call an external system are the exception, those need connectivity at the moment they fire, so a round designed for dark areas puts them at the start or makes them optional.
Yes, and it can do so per defect rather than per round. A workflow can branch on severity and asset criticality, dispatch one rectification form per defect row with the recipient and due date resolved from that row, put a timer and an escalation on each one, and require photographic evidence before closure is approved. If the work order itself needs to exist in your CMMS or ERP, the same workflow can create it there through an API call, a webhook or a structured connector rather than only inside IndeForms.
On the mobile app, yes, within limits we would rather state up front. A rule can run on-device text recognition against a photograph taken in a camera question and extract a matching value into a field using a pattern you define, useful for nameplates, serial numbers and printed labels. It runs entirely on the device, so it works offline and no image leaves the handset for that step. The known limitation is that the on-device recogniser is trained on standard fonts and does not reliably read seven-segment displays such as digital meters and scales; for those, capture the photograph as evidence and type the reading. This feature is mobile-only today and is not available in the browser client.
Yes. IndeForms deploys on-premise, including fully air-gapped installations for sites that cannot allow outbound connectivity, with each customer’s data in its own database. In that mode no vendor code runs inside your install and configuration updates are applied through signed offline bundles. Where you enable AI-assisted features, the platform never sends real readings or answer values to a model, form structure and aggregate statistics only, and any feature that inherently needs answer content runs only against a self-hosted model inside your own environment.