If you have been chatting with ChatGPT and a brand suddenly showed up inside the answer, that was not a glitch. It was an ad. Moneycontrol says nearly two months after the format debuted in India, the people paying for it are curious and still not opening the wallet wide. The line that should stop a founder mid-scroll is the price. One marketer told the paper that ChatGPT ads, launched on 27 August, are converting at less than 1.5 percent, with a cost per click of $200 to $400 in some categories.
Put that next to the same marketer's comparison and the gap is not subtle. On Meta and Google, cost per click can be as low as Rs 3 to Rs 5, with conversion rates reaching 3 percent. Organic searches on AI platforms, especially for products sold online, can convert as high as 8 percent. So the unpaid AI search is, in that marketer's telling, beating the paid ChatGPT placement. That is the whole mood of this story. Interest is real. The media plan is not moving yet.
Who actually showed up
Moneycontrol's deck line names Tanishq, Tata Neu, ixigo and Wakefit as brands that are testing. The body is more careful about who did what. ChatGPT ads went live in India with more than 50 brands at launch, according to Venu Madhav, general manager at Boopin. He also said the format remains unproven, and that the platform has not seen the traction advertisers give Google and Meta.
September is where the examples get specific. Against a reported starting base of around 50 advertisers that month, Pankaj Srivastava, founder and CEO of UnoSearch, said the later list adds useful visibility into adoption. Viren Inaniyan, co-founder and CEO of Trucommerce, said participation goes past the names mentioned at launch. In September, Libas promoted a Ganpati sale. The Man Company advertised three perfumes. ixigo promoted first-booking flight offers. Wakefit targeted people setting up their first home. AbhiBus highlighted safety information and booking support. Traya and Tata Neu are named as participants in the advertising ecosystem.
Tanishq is in the story's summary line, not in a paragraph that says what the jeweller ran. Do not invent a campaign for it. A name on a test list is not a case study.
A pilot is not a budget line
Inaniyan's sentence is the one Indian brand managers should tape above the slide. "The real traction metric is whether advertisers return and increase spending after their first campaign — not how many announce a pilot." Fifty brands at launch is a press number. Coming back in month three with a bigger spend is the business number. Moneycontrol does not report that return. It reports wait-and-watch.
Divyat Kanyal, founder of Tall Creative Studios, put the discomfort in plain language. Indian advertisers are used to Google and Meta, where targeting, campaign performance and conversion metrics are established. With ChatGPT, brands are still working out whether an ad placed inside a conversation actually becomes a customer. The ads are matched to the context of what you are talking about, not only to a keyword you typed. That is a new shelf. It is also a shelf nobody has a trusted ruler for yet.
Globally, advertisers have raised the same two complaints Moneycontrol records: measurement that does not go much past impressions and clicks, and targeting that is less precise than the platforms they already know. If your performance team lives inside a Meta dashboard, "we matched the vibe of the chat" is not a report they can defend on Monday.
Why a Rs 500 order cannot pay a Rs 2,000 click
Ishank Joshi, CEO of MobavenueAI, is the person who translates the dollar click into Indian unit economics. Quick-commerce and e-commerce companies will struggle to justify the cost, he said. For those businesses, the first transaction might bring in Rs 500 to Rs 1,800. If acquiring that customer costs around Rs 2,000, the maths does not close. You are not "building a brand." You are paying more for the door than the customer spends inside.
A $200 click is roughly Rs 16,000 at a back-of-the-envelope rate, and $400 is double that. Moneycontrol does not print an exchange rate, so do not treat those rupee equivalents as the paper's number. Even the Rs 2,000 acquisition cost Joshi uses, which is far below a $200 click, already breaks a grocery or fashion basket. The $200 to $400 range is "in some categories," from one marketer, not a rate card for every Indian campaign. Still, "some categories" at that price is enough to keep a CFO in the experiment column.
Joshi is not writing the channel off. He said it is not for everyone, and that it can still be a strong channel because the users have high intent. If the category matches, he would rather be an early investor than a late one. The categories he points to are the ones where people already research for a long time: automobiles, real estate and travel. Some retail can fit too. A flight, a flat, or a car can absorb a costly click. A Rs 149 biryani order cannot.
What this means if you only use the chatbot
For the person on the free or Go plan, the practical change is smaller than the ad-industry panic. You may see a brand inside a conversation because the chat was about a first home, a bus ticket, or a perfume, not because ChatGPT read your UPI history. Moneycontrol does not describe a new toggle, a new price, or a promise that ads stay out of paid plans. Those would be a different story. This one is about whether Tata Neu, ixigo and Wakefit keep paying.
The comparison that should calm the "ads killed ChatGPT" takes is the organic number in the same piece. The marketer said organic searches on AI platforms can convert as high as 8 percent for products sold online, higher than the paid ChatGPT placement's under-1.5 percent. Brands showing up because someone asked a real question is a different machine from brands buying a slot. If the paid slot stays this expensive and this poorly measured, the loud part of Indian marketing will stay on Meta at Rs 3 to Rs 5 a click, and ChatGPT will stay a pilot you mention in a pitch deck.
What I would not do with this
I would not move a performance budget off Google because a launch had 50 logos. I would not quote "under 1.5 percent" as an industry average. It is one marketer, talking about ads that went live on 27 August, nearly two months on. I would not tell a D2C brand that automobiles "work" either. Joshi said the equation could be different there because research cycles are longer. He did not publish a car-brand case with a cost per lead.
What I would do is watch the second spend. Inaniyan already defined the test. If Libas, ixigo, Wakefit and the rest come back and raise the amount, the channel is becoming media. If they do one Ganpati sale and a perfume test and then go quiet, India just told OpenAI that a conversation is not yet a shop. Right now Moneycontrol's headline is the second outcome's waiting room: high cost, low conversion, wallets still mostly closed.
Comments (0)
Be the first to comment!