The Most Valuable Hours in Your Business

The job is already won when the surveyor knocks. What gets written down next decides how it ends.

Last month, this column was about the customers you already have, and the value sitting quietly in your aftercare data. This month, I want to go back to where all of that data is born. The survey.

Think about where the survey sits. By the time the surveyor knocks on the door, the job is already won in most cases. The contract is signed, the deposit is in, and the customer is waiting. Everything the business has promised now passes through a couple of hours in the property. The order is placed from it. Fabrication trusts it. The fitting team plans their day around it. Even the warranty record and the aftercare register we talked about last month start life there. And yet, in a large part of the industry, that survey is still captured on a pad (or iPad!), a few phone photos, and whatever the surveyor happens to remember on the drive back.

When something goes wrong on installation day, the frame that fouls the reveal, the colour that doesn't match, the access problem nobody planned for, the post-mortem almost always ends at the same place. It was visible at survey. Somebody saw it, or could have. It just never made it out of their head and into the paperwork. That's rarely incompetence. It's thin capture. A busy person, a small pad, and a format that throws most of the detail away.

This is where AI has quietly changed what's possible. The old fix was more forms, longer checklists that surveyors, reasonably, resented. The AI version is the opposite: capture more by writing less. A surveyor who talks through the job room by room, describing what they see and what they'd flag, can have that turned into a structured record. Sizes, cill details, access notes, the thing that needs checking with the fabricator. Not a tidy paragraph. A record with everything in its place.

What's newer still is that this is becoming glazing-specific. The first wave of AI tools were generalists, brilliant at words, blind to windows. The tools now emerging for our industry are built around how a glazing survey actually works: what should be captured for a bay versus a bifold, what an installer downstream needs to know, what turns a survey into something you can order from with confidence rather than something you decode later. The gap between "notes" and "specification" is closing, and I don't think most of the industry has noticed yet.

Here's the part that interests me as a business question rather than a technology one. A structured survey doesn't just prevent errors. It compounds. The order can go to the fabricator the same day, while every detail is fresh. Every day trimmed between survey and order comes off the lead time the salesperson promised. The order can be checked against the survey automatically, which is exactly the handoff problem I hear about most often. Anything the survey uncovers that changes the job, the extra structural work, the awkward access, gets flagged while there's still time to talk to the customer about it, rather than absorbed into the margin on installation day. A new surveyor learns faster, because your best surveyor's structured records become teaching material rather than private shorthand. And that aftercare register from last month builds itself from day one.

There's a people angle too, and it's the opposite of what surveyors might fear. Structured capture doesn't devalue the surveyor; it showcases them. The experienced eye that spots the settlement crack finally gets recorded, credited and reused, instead of evaporating into a scribbled margin note. That's a retention conversation as much as an efficiency one.

The honest limits, as ever. AI won't spot the rotten lintel; that's the surveyor's eye, and nothing here replaces it. A rushed survey badly described is still a bad survey, however clever the tool that structures it. AI can spot potential issues from the information entered at Survey, but it’s the surveyor that can decide whether the potential issue is a real issue. And the conversation in the customer's house, the reassurance that they bought from the right company, the reading of what they want, remains positively human.

If you want to test the idea without buying anything, try this. On the next survey, do the pad as normal. Then dictate the job into your phone as well. Have AI structure the recording back at the desk and compare it with the pad. That comparison usually settles the argument better than any column can. Take it even further and try SurveyPal. It’s surveying AI built exclusively for glazing.

The sale wins the job. The survey decides how it ends: on time, on margin, with a customer who'd recommend you, or as another story that starts with "it was visible at survey". Treat the survey as data capture rather than paperwork, and everything downstream; order, fabrication, install, aftercare, gets easier. It has always been the most valuable time spent in the process. We're finally getting tools that treat it that way.