Marketing’s Identity Crisis: AI Still Can’t Find the Buyer

Marketers did not have a data problem. They had an identity problem. The past decade chased every click, cookie and channel as if each signal was a customer. AI was supposed to clean up the mess, but new research suggests it may just be making the confusion faster.

The identity gap hiding in the data

The American Marketing Association’s 2026 study of 203 US B2C marketers found an ‘identity gap’. Overall, 53% of marketers rated their marketing highly effective. But split the sample by identity resolution and the story changes. Among marketers using identity resolution, 64% said they were highly effective at reaching individual prospects. Only 31% without it said the same.

Personalised campaigns showed a similar divide: 68% of identity-resolution users rated themselves highly effective, compared with 39% of non-users. Marketing works, it seems, until you ask it to recognise the same customer twice.

One Indian buyer looks like six customers

Indian CMOs should take this seriously, because India may be the hardest identity lab in the world. A single buyer can discover a product on Instagram, search reviews on Google, watch a YouTube video, ask a friend on WhatsApp, visit the brand website, abandon the cart, and finally buy in a shop three days later. Six interactions. Five platforms. One human being.

The average Martech stack may record those as six separate humans. That is why identity resolution is not a back-office data cleanup: it is the difference between seeing a person and seeing fragments.

What customers are really saying

Two-thirds of consumers globally said they had experienced personalisation that felt inaccurate or invasive. Their message, the report notes, is ‘Know me. Just don’t get creepy.’ That should be the first rule of AI marketing.

The second rule is simpler: start with the person, not the platform. A customer is not a cookie, a WhatsApp number or a device ID. She may be researching a refrigerator on Tuesday and a school for her child on Wednesday. Context and intent matter more than the channel she used.

What to do before adding another AI tool

Corporate India’s numbers make the tension clear. Deloitte found 45% of Indian consumers expect brands to anticipate their needs and communicate proactively. EY India found 77% of online shoppers strongly worry about data breaches, and 73% worry about their private information being leaked. Yet 62% have already made purchases influenced by AI recommendations.

This is not a ‘collect everything’ moment. It is a restraint moment. The smarter approach is to ask what you actually need to know to make the next interaction more useful. Before scaling AI, run this short checklist:

• Can we recognise the same buyer across the devices and platforms we already own?
• Does the next message show we remember what happened before?
• What data would genuinely improve the next step, rather than just sit in a dashboard?

Measure the person, not the impression

Impressions, clicks and engagement feel productive, but they do not guarantee revenue or loyalty. If the same customer is counted as three people, a brand may congratulate itself for acquiring customers it already had. AI can target, predict and automate, but it cannot repair a broken identity beneath it.

The question is no longer how many people you can reach, nor how precisely you can target them. It is the embarrassing one: do you actually know who you’re talking to? If the honest answer is no, don’t buy another AI tool. Go meet the customer. A real human is still the best data point a marketer has; every other signal is supporting evidence.

Source: ETBrandEquity.com