There are now dozens of apps, filters, and AI prompts that claim to tell you your colour season from a photograph. Some of them are quite good. They use real colour theory — undertone, value, chroma — and produce results that are plausible and sometimes accurate. I have clients who have used them. I have clients who arrived at the studio already convinced they were a Deep Winter, or a Light Spring, and turned out to be neither.
I want to explain why that happens. Not to dismiss AI tools — they have genuine value as an entry point into the subject — but because the gap between what a photograph can reveal and what draping reveals is real, and it is worth understanding before you decide whether to act on a free result.
The problem is light.
Personal colour analysis is a discipline of observation in neutral daylight. "Neutral daylight" is specific: north-facing, indirect, without the warm cast of direct sun or the fluorescent skew of interior lighting. That light is the medium the analysis works in. Without it, you are looking at something else.
A photograph is not taken in neutral daylight. It is taken in whatever light happened to be present at the time, processed by a camera sensor with its own colour biases, edited — often automatically — by the phone's image software, displayed on a screen calibrated to look good rather than to show accurate colour, and then analysed by software trying to compensate for all of the above.
What a drape does, held against the face in north-facing daylight, is different in kind. It changes the light environment around the face. It shows the skin as it responds to a cool or warm ground. It reveals whether the jawline lifts or drops, whether the eyes sharpen or flatten, whether the skin reads clear or tired. These things are visible in person and almost entirely invisible through a screen.
The problem is also skin.
Skin is not a flat surface. It has depth, translucency, and a reflectance that varies by area — the inner arm reads differently from the cheek, which reads differently from the forehead. In natural daylight, a trained practitioner observes these variations continuously during a drape sequence. In a photograph, they flatten. The "skin tone" a camera records is an averaged reading; what we are actually assessing in a consultation is something more dynamic than that — the skin's response to different colour grounds, which changes in real time as the drapes are exchanged.
This is why two people with very similar selfies can sit in front of the same set of drapes and read entirely differently. The drapes do not respond to the image of the skin; they respond to the skin itself.
What AI tools are good for.
I should be clear: AI colour analysis tools have real uses. They have made the vocabulary of colour analysis considerably more accessible, and that matters. Clients who have used them arrive knowing what undertone means, having a hypothesis to test, understanding the difference between a Spring and a Summer at least in broad terms. That makes the consultation richer.
They are also genuinely useful for exploration — for trying out seasonal palettes virtually, for helping someone decide whether to invest in a consultation, for narrowing the field before an appointment. If an AI analysis confirms what you suspected, that is useful information. If it surprises you, that is a reason to check it in person.
What they are not suited to is producing a result you can dress from for twenty years, with confidence. That requires drapes, a north-facing window, and a practitioner trained to read what the drapes show.
What to do if you have had an AI analysis.
If you have used a free tool and got a result that felt right, I would not immediately dismiss it. Wear the palette for a few weeks and observe whether the recommended colours do what they are supposed to do — whether people say you look well, whether clothes feel more coherent. If they mostly do, the tool got it right. If something feels consistently off, a consultation will tell you why.
If the result felt wrong from the start, trust that instinct. Come in and we will look at it together. Bring the result; it is useful context, even when it turns out to be incorrect. The consultation does not start from scratch — it starts from wherever you are.
Rebecca Sells is a certified colour practitioner and the founder of The Colour Doc, a personal colour analysis studio in Battersea, South London. She practised as a clinical medical doctor before opening the studio in 2025.
→ Book a consultation What happens in an appointment Back to the journal