It’s Monday, I’m Nithya Sudhir. I collect words, chase patterns, and write about whatever makes me curious.
The AI Mirror Is Here.
Nobody buys a coat from a description. They buy it from a reflection.
On October 1, ChatGPT started offering one. OpenAI launched virtual try-on and Favorites globally.
A Try On button now appears on clothing and accessory listings in ChatGPT. Shoppers tap it, take or upload a selfie, and see how that exact piece might look on them. According to OpenAI's release notes, the selfie is saved for future try-ons, so there's no re-uploading for every item. Anything they like can go into a Favorites library for later.
That isn’t everything that happened the last quarter with OpenAI and shopping, right?
In August, ChatGPT introduced carousel ads that pull live prices and star ratings straight from a brand's website.
In July, Health in ChatGPT let people connect their medical records and Apple Health data, and ask questions grounded in their own health. And the traffic is following.
ChatGPT is inching closer to the moment of purchase. For DTC brands, that's a big opportunity, but only if shoppers trust what they see.
So today, we're asking two questions. Why does a real brain trust an artificial one? And how far does that trust go?
🤝 This edition is kindly brought to you by Omnisend
Omnisend added a countdown timer to its email editor, so urgency no longer depends on a third-party tool. Set the end date and time, pick the DD:HH:MM:SS format, match your brand fonts and colors, and upload a fallback image that replaces the timer once it expires. It's available on every paid plan.
Moving from another platform? Omnisend's team migrates your setup in 5 days, free, which gets you live well before BFCM.
Where Shoppers Trust AI (and where they don’t)
In a recent survey, 45% said that an AI recommendation increases their trust in a brand — almost half — and yet only 2% were comfortable handing a purchase over to AI entirely.
Source: Aware, Uncomfortable, and Still Not Buying: How Consumers Really Feel About Shopping in ChatGPT and Google, Crazyegg
Let’s understand that decision a bit better.
1. We don’t trust AI with taste
People believe AI gives better advice than humans on practical qualities and worse advice on sensory, experiential ones.
It’s called the word-of-machine effect.
Chiara Longoni and Luca Cian put it to the test. They asked people to choose between real estate properties. One was picked by a human realtor. The other was picked by a realtor's "proprietary algorithm."
When people judged the properties as practical investments, 60% chose the algorithm's pick. But when they judged them by how fun they'd be to live in, 76% went with the human's pick.
When the shopper's mindset changed, the trust in the machine flipped.
The study concluded that shoppers trust AI recommendations when the product is utilitarian, but resist them for products chosen for pleasure or experience.
However, teaming up AI with human decisions brought those trust levels up to human levels and that’s what Try-on does.
The thing is, fashion is a hedonic decision, so ideally, humans don’t trust its decision. But with Try-on, AI (here ChatGPT) isn’t the one giving an opinion on the outfit. It only renders the outfit on the shopper's body and lets them judge it for themselves.
Total control. Win-win!
What would you actually let AI do for you?
2. We resist AI that treats us as average
Every shopper walks into a fitting room with the same thought. It looks great on the model, but will it look good on me?
That doubt sits at the heart of a second barrier to trusting AI. Chiara Longoni, Andrea Bonezzi and Carey Morewedge called it uniqueness neglect, the fear that a machine treats everyone the same and can't account for what makes each person different.
And because fashion is one of the most personal purchases there is, this is an important caveat.
The good news is that this resistance disappeared when the AI was framed as personalized, or when it only supported a human's decision rather than replacing it.
With Try-on, the recommendation is no longer about an average body. It's about theirs. The feature removes both triggers of resistance at once — it personalizes and leaves the final call to the human.
Source: ChatGPT is becoming a shopping destination: the AI platform receives over 84 million shopping questions from U.S. consumers every week, Stackline
3. I vs AI
Today, we live in a world where people weigh advice from an algorithm more heavily than advice from a person.
But when people had to bet on either the algorithm's answer or their own, far fewer picked the algorithm.
Jennifer Logg, Julia Minson and Don Moore called this algorithm appreciation.
In their experiment, people guessed a man's weight from a photo. They were all given the same hint, 163 pounds. But half of them were told it came from an algorithm, and half were told it came from past participants.
Surprisingly, people moved their guesses further toward the "algorithm’s answer".
And this is where I love the Favorites feature.
When ChatGPT shows a product, it acts like that 163-pound hint. It nudges the shopper's choice in a direction. When the shopper saves the product to Favorites, they're agreeing with the nudge, but on their own terms.
It keeps the AI on the side people trust, as an advisor, and away from the side they resist, a replacement for their own judgment.
Machines get punished harder
On the flip side, though, people lose confidence in an algorithm faster than in a human after seeing both make the same mistake.
Dietvorst, Simmons and Massey called this algorithm aversion.
In one task, human forecasters made nearly twice as much error as the algorithm. Even so, people still leaned toward the human.
In the end, we forgive people for being wrong, but we expect machines to be right.
How DTC brands can use ChatGPT right now
Albertsons isn't a DTC brand, but it offers the clearest model for how DTC brands should treat ChatGPT.
Its new Safeway experience in ChatGPT lets shoppers start with a recipe, a photo, a digital list or a simple request like "Plan pizza night for four."
ChatGPT surfaces relevant products and savings, helps build the cart, then sends shoppers to Safeway to check out.
Why it works?
It plays to AI's strengths. Planning a meal, finding ingredients and spotting savings are practical tasks. That's exactly where shoppers trust AI.
It personalizes. The cart starts from the shopper's own recipe, list or request, which answers "but what about me?"
It keeps the human in charge. The shopper reviews the cart and checks out on their own.
It's timed right. Shoppers are more open to using AI for groceries than for other purchases.
Walmart follows the same playbook. Its in-ChatGPT app takes shoppers from discovery into a tailored Walmart experience with account linking, loyalty and Walmart payments.
The bottom line
Shoppers are happy to let AI search, compare and show them how things look. They just want the last word.
So give the machine the facts and give the shopper the final say. The brands that do both will look good in the mirror and even better at checkout.
What do you think about the new feature and why it works? I’d love to know your thoughts!
How's the depth of today's edition?
As always, hit reply if something in here hits home.
See you next week,
Nithya
P.S. If you want to get a case study about your own brand, reply to this email. If you’d like to reach our newsletter audience (founders, creators, and marketers), click the button below.
If you’re new here, I’m over the moon you’ve joined us! To help me craft content that’s actually useful (and not just noise in your inbox), I’d love it if you took 1 minute to answer this quick survey below. Your insights help shape everything I write.
Insane Media is more than one voice
💡 Dive into our other newsletters - where psychology meets the founders, creator economy, Human resources and AI trends.
