If you're like me, you've probably spent countless hours experimenting with TikTok's color analysis filters, desperately trying to pinpoint your ideal color palette. Despite my efforts, I've always been uncertain about which 'season' truly suits me.

I fall somewhere between spring and autumn.
With so many filters available, it often feels like each one gives a conflicting result.

I was fascinated when I stumbled upon this trending video by @kaligirly, where she explained how she identified her perfect color palette using .

With over 400,000 people saving the video (likely to attempt it themselves) and countless others applauding its ingenuity, I felt compelled to give it a shot. One commenter even mentioned it saved them $500 (the typical cost of a professional color analysis 😳).

I decided to test Kali's method...and the results were surprising. Here's what I learned (feel free to follow along and try it yourself!):
Kali's first step was to take a makeup-free selfie in natural light near a window, avoiding direct sunlight:


Kali’s next step was to open the photo in your camera roll and enter edit mode as if you were going to draw on it. Once you click the draw icon, select the color wheel at the bottom right:


After selecting the color wheel, tap on the color picker tool:

Kali then advised selecting a part of your face that best represents your skin tone. You can use the drawing tools to create a swatch and see if the color suits you:

After experimenting with different areas of my skin, I settled on this shade, which I felt matched the pinkish undertones in my complexion:


Next, return to the color wheel settings and switch to the 'sliders' tab to retrieve the color code for your skin swatch:

Here's my detailed process and the color code I extracted from my swatch:


Now, it's time to leverage ! Kali began by asking to generate a seasonal skin tone color palette based on her skin swatch:

For Kali, identified her skin tone as having warm undertones, suggesting she could align with either a spring or autumn seasonal color palette, considering additional factors like her eye and hair color.

Interestingly, also identified my undertones as warm when I posed the same question using my skin swatch color code. However, unlike Kali, suggested I likely fit into the 'spring' color palette, which 'typically features warm, peachy, or golden undertones' and complements shades like 'coral, peach, golden yellow, warm greens, and soft blues.'

I was surprised I didn't receive 'spring or autumn' like Kali did! I've always debated whether I lean more toward spring or autumn palettes. The peachy undertone makes sense, but I've often wondered if corals and peaches might accentuate my rosacea. Other filters usually categorize me as autumn.
Since mentioned that hair and eye color could influence her palette, Kali performed the same swatch test to determine her eye and hair color codes, which more clearly indicated she belongs to the autumn color palette:


Although identified me as a spring based on my skin tone swatch, I decided to also analyze my eye and hair colors, especially since I wasn't entirely convinced about being a spring:


And, I felt like I was back to square one. indicated I might fit into either the spring or autumn palette when considering my hazel-dark green eyes and bronde hair.

recommended I assess whether softer, lighter shades like peach, coral, and golden yellow or deeper tones like terracotta, olive green, and burnt orange suit me better to determine my ideal palette.
So...I conducted a DIY color analysis using clothes from my closet (unfortunately, I lacked the specialized tools professionals use). Since I didn't have exact colors like terracotta, I compared 'softer, lighter' shades (left column) with 'deeper tones' (right column).

Honestly, this was a lot of fun to create, but it left me a bit more confused. I think I naturally gravitate toward autumn tones, but it might depend on whether I want a softer appearance or something with more contrast. It could also vary based on how light or dark my hair and skin are, which can change with sunlight, lighting, hair dye, and the seasons. My mom sees me as a spring, but she thinks I view myself as an autumn. Honestly, this might just confirm that I fall somewhere in between (and you might too)!
Since Kali wanted to add more color to her wardrobe, she also asked for specific shades of each color that would suit her best, receiving a detailed breakdown of ideal hues for her palette.

She mentioned, 'I have to admit, it's quite accurate because it suggested coral pink, and one of my favorite T-shirts happens to be coral pink.' Kali shared that she had been curious about her color palette ever since she got this shirt, and essentially confirmed that she should incorporate more coral pinks into her wardrobe.

While didn't provide a detailed list for every color, it did suggest that I opt for peachy pinks to match my warm, peachy undertones and steer clear of cool-toned pinks with blue undertones.

...which I found quite accurate after conducting my own color test as well! Confirming that I have warm undertones was one of the most helpful takeaways for me. Until now, I wasn't entirely sure if I leaned cooler or warmer. I always wondered why purples seemed to wash me out... Now I understand that warm pinks and corals enhance the peachy undertones in my skin!

After Kali received her detailed list of shades for each color, she wanted to know which colors she should prioritize. So, she asked to rank the colors that suit her best, from most to least flattering. Here are her results:


Here are my results, which are somewhat similar to Kali's, but peachy pinks rank as my #1 (likely due to my peachy skin tone) instead of red, which is my #4 (and not a color I naturally prefer). Essentially, warm tones work best for me.

After a few years, I finally understand why this peachy-pink dress looked so flattering on me!

That wraps up Kali's AI color palette experiment (and my attempt)! While Kali believes this doesn't replace the 'precise expertise' of a professional color analyst (I agree — check out a professional analysis here), it did help her make smarter shopping decisions.
She shared with Mytour, 'Like many, my wardrobe was mostly neutral tones like white, black, grey, and brown, but I wanted to incorporate more colors. However, I struggled because trendy pastels, especially blue (my favorite), never seemed to suit me. This hack was incredibly helpful in steering me toward colors that complement me better.'

Kali found 's ranking of her most to least flattering colors particularly useful. She explained, 'I usually lean toward cool tones when shopping, but warm tones, like red, were ranked as my #1 most flattering color. Ironically, I don’t own any red, orange, or yellow clothing, so it encouraged me to explore shades I’d never considered, such as terracotta.'
She added, ' even pointed out that pastel blue is one of the least flattering colors for me, which finally explains why I struggled so much with color choices in the past.
I also had a vague sense that light blues and purples didn’t quite suit me, but this hack clarified why — cool tones just don’t work for me! After trying this, I’m definitely less inclined to buy those colors. Now, I’m on the lookout for peachy pinks and corals for summer.

I was still skeptical about 's accuracy and generalizations, but Kali mentioned that people who tried the hack and had professional color analyses reported similar results from . She said, 'My mom is one of them! She was professionally analyzed as a cool summer, and gave the same answer.'
While working on this post, my partner (who's pursuing his PhD in artificial intelligence) peeked over to see what I was doing (BTW, I think he's a summer 🤪). When I asked if this was credible or if could genuinely provide accurate advice, he said the most convincing evidence would be if people with professional analyses received similar results.

Thank you, Jarrid, for your input. 🫡
He explained, 'These language models are designed to produce human-like responses, but that doesn’t always mean they’re factual, true, or accurate.'
He added, 'The model was trained on internet data, so if there’s information about color palettes online, it might use that. However, it’s not specifically trained to provide accurate color palette recommendations; it’s trained to generate human-like answers. So while it might get it right sometimes, it could also be wrong, even if it sounds convincing. The best way to confirm is through a professional analysis.'
In short, treat AI language model outputs with caution! While I might not invest in a professional analysis, I still see value in consulting an expert (and might have some inaccuracies). That said, I’m satisfied with what I’ve learned. provided a useful framework, but my own experimentation helped me determine what works best. I also recognize that color analysis isn’t an exact science; it’s subjective and can vary based on personal preferences or even lighting conditions!
So, if color analysis tools work for you, that’s fantastic (they work for me too)! But if they suggest red looks best on you and your favorite color is blue, that doesn’t mean you can’t wear blue. Ultimately, the colors you choose to wear are entirely up to you!
