Last month, Netflix quietly deleted all the remaining user reviews on its service. And just like that, CEO Reed Hastings completed the latest shift in his company’s ever evolving strategy: away from the wisdom of crowds in its content recommendations, in favor of the wisdom of the machine. The algorithm is now firmly in charge at Netflix—as it increasingly is across the economy.
The past year has seen a crescendo of fascination with—and fear of—artificial intelligence. This theme is a central feature of the tech backlash buffeting formerly bulletproof companies like Facebook and Google, fueling U.S. congressional hearings and editorial-page outrage. Yet it’s also the most potent new engine for business efficiency, an essential driver of competitive advantage from Amazon to Accenture.
The implications of AI and the rise of the algorithm have never been more relevant, more important, or more complex. What is irrefutable: Tech’s newest wave is upon us, and there’s no going back.
Whether you run a big public company or a startup, whether you are inside academia or in the nonprofit world, whether you oversee a team or contribute to the operations of an enterprise, you need to understand AI’s emergence. Here are five lessons that underscore how the algorithm is redefining business and organizational leadership—including the biggest lesson every leader and businessperson needs to grapple with to thrive in this era.
1. Spotify: Commit to the algorithm
From an unlikely location—Stockholm—streaming service Spotify has upended the global music industry. But how? I spent time recently at Spotify’s headquarters, as well as in its New York offices, getting to know the company leadership. Many factors have fed Spotify’s success, but a key element has been the company’s commitment to technology. The simplest evidence: Some 40% of its employees are dedicated to engineering and software. CEO Daniel Ek loves music—he even spent a year trying to become a professional guitarist—but he understands that what makes Spotify special is the way its technology interacts with music, artists, and listeners. There are so many songs, and so many music tastes, that without cutting-edge software, it’s impossible to sort and prioritize all the options.
Apple CEO Tim Cook, when I met with him early this year, critiqued those who make music “about the bits and bytes.” But Spotify’s algorithm layer and its overall software prowess put it on top and give it an edge, even as Cook’s company has taken Spotify on with the more human-curated Apple Music. In the process, Spotify hasn’t done too badly for itself, generating a $30-plus billion market value.
2. Biohub: Turn data overload into opportunity
Many organizations appreciate the power of data, but getting meaning from vast amounts of information is a tricky art. Sensors and stats mean nothing without the tools to put the data they collect into context.
I got a compelling visual representation of this during a recent visit to the offices-cum-laboratory of the BioHub in San Francisco. Joe Derisi, one of the BioHub’s co-directors, showed me the output of a new kind of microscope. If a standard camera has one lens to focus, Derisi explained, this one had 22 lenses—each of which has to be finely tuned in relation to the other, in minute calibrations that would be impossible without software.
The day before, I had sat in on an advisory board meeting for a major scientific research effort. I learned how a single two-day experiment can now yield millions of detailed results, thanks to our ever-growing ability to generate and collect data. Yet it can then take nine months or more for researchers to make sense of it all. Which is why academic researchers will try to tap an engineering expert to write algorithms that speed things up. Because the faster they can make sense of an experiment, the faster they can move onto the next one, fine-tuning it to yield more powerful results.
This potential holds true for almost every enterprise today, not just in the realm of medical research. To maintain a competitive edge, you need to test, study, and evolve. Yesterday’s edge of innovation quickly becomes table stakes. Or as one CEO I spoke with acknowledged, the goalposts just keep moving. Algorithms can be central to all parts of the test-study-evolve process, but it is in that “study” phase—where opportunities can suddenly present themselves—that so many operations get bogged down.
3. Goldman Sachs: Let AI lead . . .
Two years ago, Wall Street investment bank Goldman Sachs launched Marcus, a consumer-facing retail bank for a next-generation customer. By building a tech platform from scratch (and leveraging Goldman’s balance sheet), Marcus has been able to offer higher rates of returns and better services than many traditional competitors. But as it turns out, that’s only the beginning. This spring, Marcus acquired Clarity Money, an AI-infused interaction platform that, rebranded, will become core to the Marcus experience.
Rather than being afraid of fouling up its strong start, Marcus is acting to disrupt itself, exploring a next-phase vision of its product. This is the opportunity—and the peril—for many companies. As AI gets better and better, are you prepared to embrace what it has to offer, or will you risk ceding a competitive advantage to another player?
4. Facebook: . . . and keep your head up
Facebook has ridden the speed and scale advantages of AI and the algorithm to incredible heights, but the last year has also revealed the risks in that approach. The company did not appreciate the dysfunction that those who were gaming its software could produce. If one-tenth of 1% of Adidas shoes come off assembly lines defective, they can be written off. But for Facebook, with 2 billion users, that sliver could mean 2 million bad actors around the world—a problem that simply can’t be tolerated.
What Facebook’s situation illuminates is that every action we take as business people will have unintended consequences. When an algorithm or an AI is overseeing those actions, it can initially mask the consequences. The more dependent you are on the machine, the harder it can be to identify and fix problems. And so even as the imperative to embrace the algorithm reaches urgent levels, the need for vigilance is paramount.
Some observers believe that government should provide such vigilance, and there is a drumbeat in U.S. Congress for regulatory action. Yet as tech advances, the burden—and responsibility—falls on private enterprise to better police its own choices. Facebook, for one, is racing to address its missteps, throwing more people—and yes, more software—at fixing the holes in its net. That may not prevent the inevitability of outside watchdogs or other intervention, but that doesn’t make it any less necessary.
5. Nike: Respect the human factor
When Nike chose to highlight Colin Kaepernick in its 30th anniversary “Just Do It” campaign and launch a product line around him, there were no blinking lights on a computer pointing the way. As powerful as algorithms are—and Nike uses them for everything from manufacturing to design to marketing—no computer code can make the truly challenging choices that face people in business every day. It was Nike CEO Mark Parker’s gut instinct, and his heart, that approved the Kaepernick program—his intuitive sense of where Nike is as a brand and where it could move in the future.
It would be foolish not to embrace the powerful potential that AI and algorithms offer, but we also need to stay disciplined about parts of our enterprises that need a human hand at the helm. Even Netflix, which deploys extensive software to tell me what shows to watch based on what I’ve watched before, makes expensive bets on creative showrunners to bring those shows to life. And what if I’d truly love experiencing something entirely new? The AI has no basis for dreaming up the next House of Cards, not to mention a game like Fortnite, which has attracted millions of young acolytes, many of whom are watching gameplay videos on Amazon’s Twitch network rather than what’s traditionally been conceived of as a “show.”
Nike’s bet on Kaepernick may prove costly or prophetic, but it is a choice at the heart of the company’s mission and purpose, and could never be delegated to computer code. A colleague recently shared this mantra: Everything in an organization that can be done by machines should be done by machines—efficiency dictates it. But everything that needs to be done by humans must be done by humans. The defining characteristics of an enterprise—those involving ethics, judgment, creativity, and compassion—require a human touch. We cannot lose sight of that, whatever the algorithm may tell us.