How Accurate Is AI Skin Undertone Analysis

What accuracy depends on

• Light: daylight by a window is the reference point; artificial yellow light shifts the result toward warm.
• White balance on the phone: some phones automatically "warm up" the image.
• Makeup: foundation and bronzer cover the skin's natural color.
• Tan: changes the depth of the complexion, but not the undertone.
• Background: a red or yellow background reflects color onto the face.

Three checks you can do at home

1. White and cream fabric next to your face: if white makes you look fresher, the undertone is likely cool; if cream looks more natural, it's warm.
2. Gold and silver: jewelry that "sits well" on your skin reveals the direction of your undertone.
3. Veins on the inner side of your arm in daylight: bluish leans cool, greenish leans warm.

If all three tests give the same answer as the report, the assessment is reliable. If they differ, you're probably neutral — a mixed type where both warm and cool palettes work in moderate shades. More on this in foundation by skin undertone.

When AI most often gets it wrong

• A photo taken under a ceiling light in the evening — almost always produces a "warm" result.
• An image with a filter or "beauty" mode applied.
• Heavy bronzer contouring that changes the color of the cheeks.
• A very dark or overly contrasted image where the skin isn't visible in mid-tones.

The fastest way to boost accuracy: the same pose, twice — once by the window, once under a lamp. Compare the results and keep the one taken by the window.