By Amardeep Singh
As per research, 94% of business buyers now use AI during the buying process. At times AI can generate inaccurate responses about the brand, so who will take the accountability to fix it?
Ask an AI assistant to compare the leading players in a category, and the answer arrives clean, well-organised and confident. A company may appear with a positioning it retired three years ago, beside a competitor that has exited the market. Nothing in the response signals uncertainty. The confidence is absolute, even when the accuracy is not.
This matters because AI has moved from a peripheral research tool. to a primary discovery channel. Forrester reports that 94% of business buyers now use AI during the buying process. A brand now enters consideration before a prospective buyer ever reaches its website.
Many marketers have encountered some version of this. What stands out is not the inaccuracy itself, but the absence of a correction mechanism. There is no editor to write to, no dependable route. to set the record straight. The answer simply exists, is read, and shapes a shortlist long before anyone reaches a website.
The real shift here is structural. For two decades, marketing optimised towards a destination: get the ranking, earn the click, own what happens after. That model assumed the brand’s own page was the first real conversation. Increasingly, it is the second, or third. As AI overviews and assistants become part of ordinary research, the first meaningful encounter may happen somewhere we do not control and cannot directly edit.
So, the candid answer to who fixes it is not one team alone, and not overnight. It requires a new kind of coordination.
A brand’s website is now only one part of the information environment from which AI systems form an answer. The wider record may include old interviews, outdated directories and comparison pieces written without the brand’s input. When that record is thin, dated or contradictory, the answer can reflect what is most visible and consistent, not necessarily what is most current. In India, the challenge can be sharper in newer or rapidly evolving sectors and Indian-language contexts, where credible third-party information may be less extensive.
This is why the emerging vocabulary of answer engine optimisation and generative engine optimisation, helps name this shift, but it is not the complete answer. It points to something real: information must now be legible to a machine reader, and the brand must be citation-worthy, not merely present. That legibility is not a bolt-on. It comes from sound search foundations, published expertise that says something, and independent third-party credibility. Those functions have historically sat in different rooms, with different budgets and reporting lines.
The honest reality is that nobody can guarantee what any AI model will say. Feedback mechanisms exist on some platforms, but there is no universal correction route, and no certainty that a fix in one model will carry across to another. What brands can influence is the quality and consistency of the public record these systems encounter.
The better question, then, is not who fixes it, but who owns it. Search, content and communications all have a role, but someone must be accountable for the integrity of the whole public record.
That accountability gap is also an opportunity. For marketers, the task is no longer simply to manage what a brand says, but to build what is reliably known about it: accurate information, meaningful expertise, and credible third-party validation that both people and machines can understand and trust.
The brands that take ownership of this will be the ones that remain citation-worthy in an AI-mediated world.
The author is co-founder & president, Gutenberg.
Disclaimer: The views expressed are the author’s own and do not reflect the official policy or position of Financial Express.