Reset Before You Run: The quiet crisis in AI-driven marketing

AI didn't change the marketing game. It just gave us better tools to play it badly.

Last Tuesday, I watched a marketing director explain to his team how their new AI tool would "revolutionise customer engagement." The tool, he said with the solemnity of a man announcing the cure for cancer, could generate 500 personalised emails in under three minutes. His team nodded with the enthusiasm of people being told they'd now be reporting to an algorithm named Gary.

Yet, nobody asked the obvious question: If our emails are so forgettable, then instead of sending 500 of them, shouldn't we be asking why we're sending emails at all?

Somewhere along the way, marketers got seduced by a promise more delusional than thinking quinoa would change my life. AI, they whispered, would solve everything. Strategy on demand. Campaigns that write themselves. Success without the inconvenient burden of actually understanding your customers. It's like being promised you can cook like a Michelin chef by microwaving frozen dinners. Technically you're using heat, but let's not kid ourselves about what's actually happening here.

And here's what nobody mentions at those shiny conferences: AI didn't change the marketing game. It just gave us better tools to play it badly.

We're watching teams pour budgets bigger than small countries' GDP into AI tools that optimise content nobody remembers, personalise journeys nobody finishes, and target people with the precision of a drunken dart player. It's a beautiful, expensive frenzy, impressive on the surface, but with all the depth of a puddle after a light drizzle.

It reminds me of something my photography course leader at university said a few years ago. Commenting on the camera on every smartphone and the lack of "mystery" in modern image-making, his position to us was simple yet profound: What is the role of a photographer, he asked, when everyone is a photographer? And the question is valid just as well today for AI.

We've confused access with artistry. We've mistaken automation for intelligence. And worst of all, we've forgotten that real marketing, the kind that actually moves people, doesn't start with algorithms. It starts with giving a damn about who you're talking to.

The irony is delicious. Right now, AI in marketing is everywhere. Open-source models. Automated copy that reads like it was written by a committee of robots having a particularly uninspired day. Predictive analytics that predict everything except whether anyone actually cares. Everyone's in the pool, splashing around with the same inflatable unicorns.

But when everyone has access to the same tools, the advantage evaporates faster than your dignity after a karaoke night. As David Wingate and his co-authors argued in MIT Sloan Review, when everyone has access to the same AI technology, "it may move the market as a whole but will not uniquely advantage anyone." It's powerful, sure. But so is a cannon, and you wouldn't use it to swat flies.

Because the real problem isn't the technology. It's what we're feeding it.

We're feeding our shiny new AI stacks data that's been scraped, borrowed, and filtered through more assumptions than a family WhatsApp group. Cookies that crumble, platform data we'll never own, behavioural guesses based on digital tea leaves. And when you look at what we're basing our strategies on, the picture gets even grimmer. The numbers are staggering, and not in a good way. Sixty-eight per cent of marketers still rely heavily on third-party data, the equivalent of asking a stranger's opinion about your spouse's preferences. Fifty-three per cent admit their personalisation strategies lack accuracy, which is corporate speak for "we're guessing, but with confidence."

This isn't a performance issue. It's an existential crisis in a Canva presentation.

Meanwhile, we've spent years letting platforms like Google and Meta play matchmaker with our customers. We've outsourced our relationships to companies whose business model is literally to insert themselves between us and the people we're trying to reach. It's like hiring your ex to set you up on dates.

So before we plug in the next AI miracle cure, maybe it's time for what your therapist would call "a moment of radical honesty."

Who are your customers? Really. Not what the dashboard says, not what the lookalike model suggests, but what do you actually know about them that couldn't be scraped from their browser history? What signals are they giving you that aren't filtered through an algorithm designed by people who've no clue of what your specific needs are?

Here's the uncomfortable truth: most brands can't answer that question. We've become so dependent on digital intermediaries that we've forgotten how to have direct conversations with the people who pay our salaries.

This is where first-party data stops being a buzzword and becomes a lifeline. Not just because of privacy regulations or platform changes, but because it's the only path back to actual differentiation. If AI is the engine, first-party data is the fuel. Without it, we're all just running on premium-priced fumes.

But owning data isn't enough. You need participation. Active engagement. Customers who choose to tell you things because you've given them a reason to care. This kind of intelligence can't be scraped, purchased, or inferred. It has to be earned, one conversation at a time.

That means creating moments instead of messages. Invitations instead of impressions. Whether it's interactive content that rewards curiosity, tools that solve actual problems, or experiences that make participation feel like play rather than work, the principle remains the same: make people want to engage, not because you've targeted them, but because you've intrigued them.

In this model, AI has a role, a significant one. But it's not pulling marketing miracles from digital hats. It's the skilled assistant. The translator of human signals into actionable insights. The amplifier of what you already know, not a substitute for knowing anything at all.

The average attention span online is eight seconds. The average campaign approval cycle is three weeks. Something's not adding up. We're spending more time perfecting outputs than understanding inputs, optimising headlines nobody reads instead of fixing strategies nobody remembers.

We don't need more content. We need more context. And context begins with resetting how we think about data, participation, and what we're actually offering people in exchange for their time.

Because if marketing is about relationships, and relationships are built on trust, then trust starts with being trustworthy. Not clever. Not optimised. Not artificially intelligent. Just honest about what you know, what you don't, and what you're willing to give in return.

In a world where everyone's using the same tools, real advantage comes not from how well you automate, but from how deeply you connect. The future won't belong to those who sound the smartest or generate the most content. It'll belong to those who remember that behind every data point is a person who chose to share something with you.

So by all means, use AI. Embrace the technology. Build the most sophisticated marketing stack your budget can handle. But build it on solid rock, not rented land.

First-party data is that rock. Everything else is just expensive distraction.

(Trilokjit Sengupta is Co-founder of WNNR.in, a gamified engagement platform that helps brands capture first-party data through interactive experiences.)

Views are personal and do not represent the stand of this publication.