You Can't Beat AI With AI — You Can Only Out-Human It

A while back, I heard a story about buying a pair of diamond earrings.

It's about Lokesh Alluri, Senior Manager of Digital and Customer Analytics at Lenovo. He wanted to buy a pair of diamond earrings for a family member — a big-ticket purchase, over $2,000. By old habit, for something with this many moving parts and specs, he would have spent weeks grinding through research.

This time, he opened ChatGPT first.

What is clarity? How is cut graded? Which certificates should you look at? After a string of questions like these, he walked into the store with a working, well-informed sense of what he was buying.

Weeks of homework, compressed into days. By the time he stepped through the door, he said, he was an educated buyer.

Wow. It looks like a small thing, but behind it hides a change that no one in marketing can escape.

What Is Personalization?

Start with his employer. Many people know Lenovo because of ThinkPad. But this company's scope goes far beyond laptops: it does business in about 180 markets, sells direct-to-consumer online, and its product lineup runs from Motorola phones and tablets all the way to enterprise servers.

At that scale, what counts as personalization?

Lokesh offered a beautifully simple definition: Turn every click into a good experience.

Keep customers from getting lost in complicated product decisions, keep the digital journey smooth and free of detours, and stay relevant without crossing the line.

Notice that last part: crossing the line.

Personalization today, he said, is the customer's default expectation — every customer assumes you already understand them. So the real dividing line is this: how do you personalize so it feels helpful rather than invasive? At the same time, privacy and compliance, both on and off your site, matter more than ever.

Let's be blunt: stuffing "Dear [First Name]" into an email is not personalization. That's called laziness.

Analysts: Think More, Do Less

So with AI here, what's left for analysts to do?

Lokesh's judgment: going forward, analysts will spend more and more of their time thinking; the doing, AI will take over — and it keeps getting faster at it.

Pulling data, producing reports — the value of that kind of work is shrinking. He holds his work to one standard: can this report support a concrete decision? Can it push things one step in the right direction? If yes, it has value; if not, it's just busywork.

Ajit Sivadasan, former President of Global E-commerce at Lenovo, once said a line Lokesh has never forgotten:

You can't beat AI with AI. You can only be more human than AI.

What a line.

That sentence also shapes what he looks for when hiring. Faced with a stack of résumés that all list "proficient in AI tools," he looks for only two things: curiosity and critical thinking.

A candidate uses ChatGPT or Claude on a written test? He doesn't care. What he wants to see is whether the person genuinely understands the problem behind the question. Everyone has the tools; understanding can't be faked.

Your Funnel Is Being Folded

Back to those diamond earrings.

Before large language models, a purchase decision above $2,000 meant a long stretch of comparing, searching, and waiting. Now? The customer finishes the homework with AI first and arrives at your website already well-informed, ready to buy at any moment.

The traditional funnel has been folded in half.

And that raises a very practical problem: the conversations your customer has inside AI — you can't see them.

Today, when a customer arrives through search, you can still analyze keywords in Google Ads and guess roughly what they want. But what about the customer who arrives from ChatGPT, from Claude? In your data, there may be nothing left but a lone, clinical entry: "traffic from some AI."

That's the newest missing piece of the customer journey map.

As these AI platforms start running ads and making recommendations, Lokesh predicts, new mechanisms will emerge that at least hand back partial signals — what the customer's intent is, what they're looking for. Even if you never get the full original prompts.

And wiring those signals into your systems means major surgery. In his words, the martech stack needs "massive architectural rework" to make room for AI-mediated customer behavior.

When Everyone Has AI, What Do You Compete On?

There's one more question, and it may sting even more.

When the best technology and the best AI are things everyone can buy, what do brands compete on?

His answer: Brands will start competing on who is more human. Whoever tells their real, core story wins.

Technology will level out. Stories won't.

One Last Thing

I've kept coming back to that diamond-earring story since.

The journey is changing — customers walk in already holding the answer. The map is changing — AI conversations have become a blind spot. And people? Their value is actually getting clearer: from making reports to making judgment calls.

When change comes this fast, what do you do?

Perhaps the answer was written in that line all along: don't try to beat AI. Find a way to be more human than AI.

Here's hoping that in this round of change, you think it through first, then act.