Relationships Matter: The Human Calculus No Algorithm Can Replace

What happens when human connection outperforms a $7 million incentive program?

The portly man ran his fingers through a graying circle beard, paused, then smiled. "I am prepared to use any tool that does not run contrary to the Ten Commandments to be successful."

With that, Evgeny extended his hand, and we found ourselves in "business" with a Russian oligarch.

I remember thinking how surreal this felt, negotiating with a man whose business methods I could only guess, all in service of optimizing an algorithm. I definitely didn’t learn that in business school.

It had started a few days prior, almost like a joke, when a Pakistani product manager, Korean designer, and I, a Mexican user-researcher, gathered with Evgeny's team in an old Soviet hangar on the outskirts of Moscow. Uber had sent us there to better understand and improve our seven-million-dollar incentive program for fleet partners, who accounted for thirty percent of global trips.

What we discovered wasn’t just a taxi operation, but a vertically integrated empire. Evgeny’s ecosystem kept 10,000 drivers running across 16 cities, with a singular objective: keep vehicles in operation 24/7, 365 days a year.

The deal was simple: we'd pay Evgeny a cash bonus for incremental trips. But our riskiest variable wasn't price, it was people. How would he motivate 10,000 drivers to work harder? Our trips-for-cash deal was with the fleet, but what deal would he make with his drivers?

I expected sophisticated incentive structures, perhaps a clever bonus system that might revolutionize our thinking. What we found instead was simpler and more profound.

***

The next evening we camped out in that cavernous warehouse, watching Evgeny's dispatch team in action. The room was dimly lit, dispatchers hunched over desktop computers, faces illuminated by blue screen light. No sophisticated tracking systems, no leaderboards, just people, phones, and well-worn desk chairs.

As dawn approached, when driver activity typically declined, a dispatcher, cigarette dangling from his lips, picks up the phone, calls a driver, and simply says, "Come on, do me a favor, man. Just one more ride."

And it works. No cash needed. That simple interaction revealed a powerful truth: a human asking for a favor in the early-morning hours is far more compelling than any cash bonus our algorithms could deliver.

While AI can predict behavior, it can't mimic the relational calculus that determines whether someone will go the extra mile. Relationships matter.

***

Lately, as my job becomes more and more about automation, robotics, and leveraging AI, I’ve been spending a lot of time thinking about what actually makes work meaningful. And I keep coming back to this experience.

What struck me about Evgeny's operation was its simplicity. In a world obsessed with optimization algorithms, the most powerful tool in that warehouse was a relationship that no AI could replicate. Perhaps that's our challenge: as we build increasingly sophisticated systems, we must preserve the spaces between algorithms where the messy, complex, and profoundly human exchange of favors, trust, and connection flourish.

I don’t have an answer, to be frank. But I do believe the answer isn't choosing between human and artificial intelligence, it's about designing systems that amplify what makes us most human.

---

With gratitude to Emma Dorge and Claire Butler for the edits this week — this one is long overdue!