Every few weeks this year, a conversation with a brand has opened the same way. Our traffic is down, and we think Google is penalising AI content.

Two core updates landed in March and May; August brought a crackdown on spam and thin, mass-produced pages, and a fresh spam update began rolling out globally in the last week of September, still working its way through the index as I write. Somewhere in that sequence, the story settled into a simple and comforting explanation: the machines wrote it, so the machines were punished.

Google has not said, in this update or any other this year, that machine authorship is the thing it is acting against. The explanation does not survive contact with the evidence either. Plenty of pages produced with AI assistance held their positions perfectly well this year, and plenty of painstakingly human-written pages lost ground. What these updates consistently pushed down was not machine involvement, but sameness. Pages that had,  what several other pages in the category already had, in a slightly different order, lost visibility regardless of who or what produced them.

Which brings me to what I think is actually happening, and it is more uncomfortable than an algorithm story. AI did not create the sameness problem in brand content. It exposed one that had been there for years. A great deal of what brands published long before any of this had nothing particularly unique. It ranked, because producing content took writers time and a significant budget, and that friction quietly limited how much undifferentiated material could exist. Once the friction disappeared, everyone could produce competent, well-structured, entirely forgettable content in an afternoon. The moment everyone could, none of it was worth surfacing. Google did not change its mind about originality this year. Originality simply became the only thing left that was scarce.

This is why the question of whether AI wrote something has never been the useful one to ask. Original does not mean handwritten. It means the piece could not have come from anyone else in the category. A brand that has spent a decade in Indian retail knows things about how its customers actually buy. An insight that no model can infer and no competitor can borrow. A founder who has watched a category turn twice has a view worth disagreeing with. A marketing team that has run campaigns across fifteen states has seen patterns that never made it into any published research. All of that is original, whether it reaches the page through a writer, an interview or an AI-assisted draft built on a genuinely specific brief.

The festive quarter makes the point better than any argument I could construct. Over the next few weeks, thousands of Indian brands will publish gifting guides, fasting recipes and muhurat explainers, produced at speed so nobody is late to the season. Read fifty of them, and you will struggle to tell which brand published which. That is why an AI assistant helping a shopper decide what to buy this Diwali has no particular reason to reach for any of them. There is nothing in there that only one brand could have said. In Hindi and other Indian languages, the gap is wider still, because most vernacular festive content is translated from an English original rather than written for that audience in the first place, leaving an open field for any brand willing to do the harder thing.

The brief is where this is won or lost. A brief built from a keyword, a word count and a competitor gap will produce generic content no matter how talented the writer or how sophisticated the tool. Think of two different briefs for the same festive piece. One asks for eight hundred words on Diwali gifting ideas for young professionals, built around a keyword the whole category is already crowding into. The other asks its merchandising head which gifting categories behaved differently last Diwali from what everyone had predicted, and builds the piece around that answer. Same season, same tools, same word count, and only one of them has produced something a competitor could not have published. The tooling was never the variable. What we ask it to work from always was.

None of this makes AI the villain of 2026, and I would be wary of any agency selling that story. Used well, it removes the mechanical work that was always the least valuable part of content production. What it cannot do is supply a point of view, and this year has made clear that a point of view is the only asset in content that does not depreciate.

So it is worth reading your own festive content the way a stranger would, with the logo removed. If it could belong to any brand in your category, it would perform like any brand in your category. The brands that come through the next few updates in good shape will be the ones that had something unique to tell.

The author is managing director, NP Digital India.