The fastest way to build a reliable asset register is to validate the list you already have, not to resurvey every asset from scratch. Most FM providers already hold partial asset data from previous contracts, maintenance schedules and old surveys. Load it, take it to site, and confirm, correct or add what is there. Capture tools that only work from a blank page make you pay twice for data you already own.
AI photo capture tools are fast on a blank page
The new wave of AI photo capture tools is quick on site. A surveyor points a tablet at a data plate, the tool reads the make and model, and a record appears without anyone typing. On a new building or an estate nobody has documented, that saves real time, and any FM team rolling out a CMMS should kick the tires on what these tools can do.
They are built around one assumption, though: that every survey starts from zero. Each asset is something to discover, photograph and create. For a lot of technical services teams, that assumption is wrong on day one.
Collection and validation are two different jobs
Asset collection means discovering what exists. Asset validation means checking that what you already know is still true. Both happen on site, but they start from very different places, and most FM contracts need far more of the second than the first.
Think about a typical mobilisation. The outgoing contractor left a spreadsheet of 1,000 assets. Some are accurate, some have moved, some were replaced years ago and a few were never there. A collection tool that can't bulk-load that list and check it gives your team one option: walk the estate and photograph every asset again, including the hundreds that were already right. You lose site days, you risk duplicate records, and you throw away whatever history came with the original list.
That inherited spreadsheet is the norm, not the exception. In SFG20's 2026 report, half of nearly 200 FM professionals said they still keep at least part of their asset register in spreadsheets, and 38% either never update their register or don't know how often it gets updated. Most providers aren't starting from nothing. They're starting from a list nobody has checked in a while.
This is the exact reason one UK FM technical services team chose Ninox over AI photo tools. They needed to load the asset list they already held and check it on site, not start over.
Checking that list on site costs far less than living with it. Writing in Harvard Business Review, data quality expert Thomas C. Redman describes a rule of ten: a unit of work costs ten times as much to complete when the data behind it is flawed. Every asset confirmed during the survey is one less flawed record feeding job lines, schedules and client reports.
Validation works when the register lives in one database
Getting the register out of spreadsheets pays off on its own. The same SFG20 survey found that organisations using software reported asset register accuracy of around 70% on average, compared with around 60% for those relying on spreadsheets or paper.
Ninox is an AI-native low-code platform, so your team builds the validation process around the list you already have. It starts with a CSV import. Large files are written in batches, so a register with thousands of assets loads without timing out halfway through, and each table holds up to around 10 million records, which leaves plenty of headroom for multi-site rollouts.
Once the list is in, surveyors take it to site on tablets in the browser. Each asset already has a record, so the job becomes confirming, correcting or flagging it, with photos and signature fields on the form. Anything new gets added on the spot. Anything missing gets marked, which gives you an exception list instead of a silent gap.
Back at the office, colleagues open the same records, review the photo or data plate and complete what the surveyor didn't have time for. Teams that want data plate recognition can build it by sending the photo to an AI model through the API. And because SFG20 codes, CIBSE Guide M life expectancy and NRM codes sit in linked tables, triggers can fill in the right maintenance data when an asset type is confirmed.
None of this needs a development team. The engineers who know the assets build and adjust the database themselves on a low-code platform, and IT approves it without carrying the build.
Every wasted site visit gets more expensive
Resurveying known assets was always a poor use of an engineer's day, and it is about to cost more. BCIS forecasts that its All-in Maintenance Cost Indices will rise 15% by the second quarter of 2030. Every site day spent photographing an asset you already had on record is a day that could have gone to the assets you didn't.
When a new contract lands and mobilization is pedal to the metal, the teams that move fastest will be the ones who start from what they have. Load the register, validate it on site and build the process on a low-code platform your own engineers can change as the contract grows.
Take the next step.
Let's figure it out together.
