Home / Academy / AskBiz Tutorials / Tips for Getting Better Receipt Scan Accuracy
AskBiz TutorialsBeginner3 min read

Tips for Getting Better Receipt Scan Accuracy

Practical advice on lighting, positioning, receipt preparation, and image cropping to maximise the accuracy of the AskBiz AI receipt scanner.

Key Takeaways

  • Good lighting is the single biggest factor — natural daylight or a bright overhead light produces the clearest images.
  • Flatten and straighten the receipt before photographing to avoid distortion that lowers confidence scores.
  • Keep the camera parallel to the receipt surface and fill 80 percent of the frame with the receipt.

Get the Lighting Right

Thermal receipt paper has very low contrast — the ink is a dark grey rather than true black, and the paper background is off-white. In poor light the contrast difference almost disappears, making it hard for the AI to distinguish individual characters. To maximise contrast, use bright, even lighting. Natural daylight from a window is ideal. A bright ceiling light works well too. Avoid taking photos under warm yellow tungsten bulbs, in shadows, or under flickering fluorescent tubes. If you are outdoors and using sunlight, make sure the sun is not casting a harsh shadow across the receipt from your phone or hand.

Flatten and Arrange the Receipt

Curled, folded, or crumpled receipts create physical distortions that change the apparent shape of printed characters. Before photographing, take a few seconds to flatten the receipt. Place it face-up on a flat, dark-coloured surface — dark surfaces provide contrast behind the receipt edges, which helps the AI locate the boundaries of the document. If the receipt is very long, try to have the most important section (vendor name and total at the top and bottom) visible within the frame. You can fold the middle of a long receipt to fit both the header and footer into one photo.

Free — no card needed

See this in action for your business

AskBiz tracks these metrics automatically — just connect your data and start asking questions.

Start for free →

Frame the Receipt Correctly

Hold your phone directly above the receipt with the camera lens pointing straight down, parallel to the receipt surface. Angled shots cause keystoning — a perspective distortion that stretches text on one side. Aim to fill roughly 80 percent of the frame with the receipt, leaving a small border of the dark surface visible around all four edges. This gives the AI clear boundaries to work with. Avoid including your hands, table clutter, or other objects inside the frame. The phone viewfinder will show you exactly what the camera sees — check the preview before tapping the shutter.

Wait for Focus Before Shooting

Most phone cameras take a fraction of a second to autofocus after you reposition them. If you tap the shutter immediately after moving the phone, the image may be slightly out of focus, which causes blurry characters and low confidence scores. Watch the viewfinder and wait until the focus indicator (usually a small square or circle) stabilises before tapping the shutter. Some phones play a subtle sound or show a brief flash when focus locks. Alternatively, tap on the receipt in the viewfinder to force the camera to focus on that area, then tap the shutter.

Handling Faded or Old Receipts

Thermal receipts fade significantly over time, especially if exposed to heat, sunlight, or friction. If you are scanning an old receipt, try scanning it near a bright light source to make any remaining ink as visible as possible. If the receipt is extremely faded, consider uploading a photo taken with your phone's high-contrast or document scanning mode (available in some camera apps as a 'scan' or 'document' filter). After uploading, even a well-processed faded receipt may return low confidence scores, in which case manual entry is the more reliable option.

Related Articles

How to Scan a Receipt Using Your Phone Camera3 min · BeginnerHow the AI Receipt Scanner Extracts Data4 min · IntermediateUnderstanding AI Confidence Scores on Receipts3 min · Beginner