Turning customer notes and photos into an inventory list

A customer's message might say "three-bedroom house, normal furniture, some boxes". That's not a list you can price from. An inventory list names the items, room by room, so the quote matches the job.

AI tools can do the typing part of this. They can't see what's behind the wardrobe.

What the tools can do

Two kinds of tool matter here.

Text tools take a messy note and sort it. Paste in what the customer wrote and ask for a table with one row per item, grouped by room. The tool can also list what's unclear: "garage contents not described".

Image tools look at photos. Wikipedia lists object detection among the branches of computer vision, the field of getting computers to make sense of pictures.¹ In practice that means a tool may pick out a sofa or a fridge in a photo. It may miss things that are partly hidden, crowded together or out of frame.

Some tools also read text in images, such as a handwritten list the customer photographed. That's optical character recognition, the conversion of images of typed, handwritten or printed text into machine-encoded text.² Check any scanned handwriting against the photo.

A layout that works

Ask for the same columns every time:

• Room
• Item
• Count
• Size or notes (for example, "flat-pack" or "glass top")
• Needs checking (yes or no)

The last column matters most. Tell the tool to mark anything it guessed, and to leave a cell blank rather than fill it with a guess.

Why the list affects the price

How much the list matters depends on how you charge. Wikipedia says UK and Australian movers price by the volume of the load, not its weight.³ In the United States and Canada, long-distance cost is generally set by factors including the weight of the items to be moved.³

Either way, a missing wardrobe or a forgotten shed changes the job. A tool that quietly drops items makes a quote too low. A tool that adds items makes it too high and may lose you the booking.

What you still check by hand

Mistakes are expected. The Australian privacy regulator notes that AI systems can give false results.⁴ For inventories, that means checking these by hand.

• Count the big items yourself. Beds, sofas, wardrobes, fridges, washers.
• Look for what photos don't show. Lofts, sheds, garages, storage units, balconies.
• Check for awkward items. Pianos, safes, gym equipment, large TVs, glass, artwork.
• Ask about access. Stairs, lifts, narrow doors, long carry distances, parking.
• Confirm with the customer. Send the list back and ask them to correct it. A short call or a site visit settles the doubtful rows.
• Check scale. A photo can't tell you if a table is small or large unless something of known size sits beside it.

Customer photos are personal

Photos of someone's home can show family pictures, mail, documents or a street number. The same regulator advises keeping personal information out of public chatbots and similar tools.⁴ Before you upload photos to any tool, read what the maker says about storing and reusing them. Crop or skip images that show personal papers. Other countries have their own rules, so check your own regulator.

How to test a tool before you rely on it

Take a past job where you know the true list. Feed the tool the same notes and photos the customer sent. Compare line by line. Count what it missed, what it invented, and what it got wrong. Do this for a small flat and a crowded house, because the results will differ.

Then decide where it fits: sorting text is the simpler job, and reading photos needs a human check on every list. What happens next is covered in preparing a quote.