Industries Manufacturing
Quality checks at the line, from a photo.
Assembly, labelling and packing are checked by eye, one unit at a time. A phone, a tablet or a camera at the station takes the photo. Sight checks it against the build standard and says pass, fail or needs a person.
- Every breaker in its slot Pass · 12 of 12
- Serial matches the works order Pass · Match
- Warning label on the door Fail · Not found
- Cable ties on every loom Needs a person · Hidden by the door
- Check
- Warning label on the door
- Why
- No label found in the door photo
- Then
- Fit the label and retake the photo
An example of the checks on one unit. Yours are written with you, from your build standard.
What Sight checks in manufacturing.
Each check is a rule written with you and trained on your own photos. Every answer comes back as pass, fail or needs a person, with the photo and the reason.
Assembly complete
Every part the build standard calls for is found, counted and checked for position: this part inside that housing, this many fasteners on that plate.
Labels and serials read
Rating plates, serials and batch labels are read and compared with the works order.
Defects found
A model trained on your own photos of your own defects finds them and marks where they are. Sight is scored on photos it has never seen before it is allowed to judge anything.
Packing verified before dispatch
The count, the labels and the required photos are checked before the pallet is wrapped.
On the line with no connection Rolling out
The check engine is being rebuilt to run on the device itself, so a station with no network gets the same answer as the server.
Sight runs in production on telecommunications install photos, where the question is the same one a factory asks: is every part there, is it the right one, is it in the right place? We have not yet deployed in a factory, so the first step is a trial model on photos from one of your lines.
- 01
Label
We label your photos, with your people confirming what counts as right.
- 02
Train
We train and tune a model on them, and score it on photos it has never seen.
- 03
Set up checks
We write each check with you: what must be in the photo, and what passes.
- 04
Connect
We connect Sight to the app your crews already use and the systems behind it.
- 05
Stay
We keep running it, and retrain it as your reviewers correct it.
You hire no machine learning team and buy no infrastructure. How we deliver
Funding that may help pay for it.
No program pays for Sight itself. These are the Australian programs whose own rules cover businesses like yours adopting new technology. We are not grant advisers: each program’s own page says who qualifies.
Transforming Queensland Manufacturing Grants Program, Round 2
Queensland Government Open, closes 25 November 2026Matched grants of up to $1.5 million for innovation and advanced technology in Queensland manufacturing. The program’s page
Innovative Manufacturing Adoption Fund
NSW Government Stage 1 closes 6 October 2026; Stage 2 due early 2027Stage 1 pays $10,000 to $50,000 towards feasibility studies and pilot projects, including assessing new software or systems. Stage 2 will fund putting the technology in. The program’s page
AI Adopt Centres
Australian Government OpenFree specialist advice, training and demonstrations for smaller businesses in priority sectors. One centre focuses on manufacturing and robotics. Advice only, no cash. The program’s page
National Reconstruction Fund
National Reconstruction Fund Corporation Open, apply at any timeLoans, equity and guarantees, not grants, for Australian-based projects in its priority areas. The program’s page
R&D Tax Incentive
Australian Government OpenA tax offset for genuine research and development, open to every industry and company size. Buying finished software does not count; an experiment to build or adapt a checking method may. The program’s page
Checked 3 October 2026. Programs open and close; the program’s own page is the authority.
Bring photos from one station on one line.
An engineer will tell you what Sight can check in them. If you have no photos yet, we will tell you which ones to take.