by Marc Benioff
So I sketched out my idea on a restaurant napkin. And the very next morning, I went to our legal team and asked them to register the domain for “AppStore.com” and buy the trademark for “App Store.”
Shortly thereafter, I learned that our customers didn’t like the name “App Store.” In fact, they hated it. So I reluctantly conceded and about a year later, we introduced “AppExchange”: the first business software marketplace of its kind, and the first major initiative born out of our new commitment to seek innovation everywhere.
* * *
About two years after we launched AppExchange, I returned to Apple’s Cupertino headquarters in 2008 to watch Steve unveil the company’s next great innovation engine: the sprawling, boundaryless digital hub where millions of customers, developers, and partners could create their own applications to run on Apple devices. Steve was a master showman, and this presentation didn’t disappoint. At the climactic moment, he said four words that nearly floored me: “I give you App Store!”
All of my executives gasped. When I’d met with Steve Jobs in 2003, I already knew he was playing a hundred chess moves ahead of me. None of us could believe that Steve had landed on the same name I’d originally proposed for our business software exchange.
For me, it was exciting and humbling. And Steve had unwittingly given me an incredible opportunity to repay him for the prescient advice he’d given me five years earlier. After the presentation, I pulled him aside and told him we owned the domain and trademark for “App Store” and that we would be happy and honored to sign over the rights to him for free.
By 2019, AppExchange had more than five thousand apps available for purchase, ranging from sales engagement and project management tools to collaboration aids. And nearly 90 percent of Salesforce customers were using them.
Steve helped me understand that no great innovation in business ever happens in a vacuum. They’re all built on the backs of hundreds of smaller breakthroughs and insights—which can come from literally anywhere. Building an ecosystem is about acknowledging that the next game-changing innovation may come from a brilliant technologist and mentor based in Silicon Valley, or it may come from a novice programmer based halfway around the world. This principle applies to technology, of course, but it applies to intellectual, scientific, and theoretical breakthroughs as well.
Whether in the business of software, retail, arts and culture, or anything else, a company seeking to achieve true scale needs to seek innovation beyond its own four walls and tap into the entire universe of knowledge and creativity out there.
Innovation, Everywhere
In 2014, Salesforce purchased a company called RelateIQ, whose software captured data from users’ email, calendars, smartphone calls, and social media posts and used it to provide important insights and reminders. If a sales rep hadn’t heard back from a customer, for instance, the software would put all the relevant variables together—like the date of their initial meeting and the dates of all prior correspondence by both phone and email—and automatically generate a task reminding the sales rep to follow up. At the time, it was the closest thing we had come to building AI capabilities into our software. We knew it was the kind of practical and predictive intelligence we needed to spread across our business.
When I announced our big AI initiative a year later, one of our first projects had to do with a problem that our sales team had been describing in feedback sessions for years: They were desperate for a tool to help them prioritize their efforts so they could stop spending too much time on accounts that never bore fruit.
So a small group of data scientists led by Hernan Asorey took on the challenge, and they built a system of “opportunity scoring.” Essentially, the algorithm looked at variables such as length of time a sales opportunity had been in play, how many competitors were going after the same account, the dollar value of the account, and the account team behind it. Then it would generate an opportunity rating of one to five stars. By tracking the outcome against its own predictions, the algorithm could get smarter over time and make better suggestions in the future.
This tool was a huge hit with our sales team, who saw a productivity boost right away. So in the summer of 2016, we made opportunity scoring available to some of our customers as well. We still had quite a bit of distance to travel before that Dreamforce 2016 kickoff date, but we were on our way.
This was around the time I decided that our AI campaign needed a name. This might sound trivial, given everything else we still had to accomplish, but we couldn’t go around calling it “the AI project” forever; it needed an identity. I thought it should be something dramatic and recognizable, like IBM’s Watson. Not surprisingly, I made a plea for “Einstein.” It was too fitting to resist: “Salesforce Einstein, the world’s smartest CRM!” Plus, to build Einstein the machine, we’d have to channel the spirit of Einstein the man.
In the months leading up to Einstein’s scheduled debut at Dreamforce 2016, a small team, lead by engineers John Ball, Vitaly Gordon, and Shubha Nabar, worked around the clock from a cavernous office space beneath the West Elm furniture store in downtown Palo Alto, where they had decamped to get away from the distractions of Salesforce headquarters.
But as it turned out, even those brilliant minds in that improvised laboratory would need to seek innovation outside its four walls.
Hernan had worked at Salesforce since 2014 in the important, though not exactly flashy, role of employing data to understand how our customers use every single software product. Then he applied those learnings to guide product decisions about which new features and releases to invest the most money in, based on customer adoption and market trends.
One of the great things about embracing innovation as a core value and creating a strong ecosystem to feed it is that everybody at every level of Salesforce is encouraged to share new ideas. At Salesforce, we believe that a good idea is a good idea, period, no matter where it comes from. As a result, everyone from the summer interns to the senior management can feel confident that their ideas can work their way into the conversation.
Which is probably why, in early 2016, Hernan didn’t hesitate to approach me to let me know that we had a problem. During a management meeting he had attended a few weeks prior, I’d asked my sales executives from various regions to forecast their numbers for the quarter. Every one of them told me that they were on target with internal projections—and yet when our numbers for the quarter came in, one of our sales execs had turned out to be dead wrong, leaving us with an unforeseen shortfall.
But Hernan wasn’t just voicing his concern. He had an idea.
What if he could feed all the quarterly sales data into Einstein and build a financial forecasting tool that could get smarter and smarter as time went on? If he got the algorithms right, he figured, we would no longer need to rely on our executives to make accurate quarterly forecasts. AI would do it for us.
This was a big undertaking. And to execute it, Hernan would need to be freed up from his usual duties. This wasn’t ideal, but we knew that if we were going to crack the code on AI, innovation would have to take priority over everyday products and functions. So Hernan requested and received time and resources from his boss, Alex Dayon.
It was the trailblazing spirit at its best, and Hernan pulled it off. These days, when I have my twice-monthly meeting with twenty of our top executives, Einstein is always in the room. After my executives offer their opinions and predictions about different regions, products, and opportunities, I turn to the virtual Einstein on my phone to see what he thinks.
Einstein then gives me its over/under prediction for the quarter, identifies where we’re strong and weak, and even points out specific patterns or areas of concern. Sometimes Einstein’s analysis can be painful to hear, and I’m sensitive to that. But it’s rarely wrong. And beyond that, it has given us something I believe every company needs: an obje
ctive, unbiased, unemotional voice at the table.
Even though it was outside his job description, Hernan developed something that ultimately became more than just an internal tool for our sales executives. It became Einstein Forecasting, one of our hottest products. In fact, after I described it during an interview with Jim Cramer on CNBC, my phone began blowing up with calls from CEOs of companies big and small. Every one of them wanted to know if they could have Einstein Forecasting.
On September 19, 2016, we formally introduced Salesforce Einstein at Dreamforce, right on schedule. As proud as we were of this achievement, we were also aware that our work was far from done. This was the beginning of our AI journey, and we would need the full force of our ecosystem to make more progress.
A Work in Progress
One summer evening in 2018, tucked into a booth at a downtown San Francisco restaurant called Boulevard, two of Salesforce’s brightest technical minds ordered a bottle of red wine and launched into one of their regular monthly strategy sessions.
This one would be devoted almost entirely to their favorite topic, AI.
Bret Taylor, our new chief product officer, and Richard Socher, our chief scientist, were already many months into the next phase of our AI initiative. Among other things, they were working building apps with deep learning capabilities that could understand anyone, no matter what language is spoken, and digital assistants that could converse, in voice and chat, in anyone’s natural language. Their session on this evening was the result of ambitious acquisitions, led by John Somorjai and his team at Salesforce Ventures, the venture arm we founded in 2009 to keep an eye on emerging trends and promising start-ups as well as to extend our innovation ecosystem, fill gaps in our product technology, and find talented computer scientists just like Bret and Richard to bring in.
Bret, who is full of energy and wisdom beyond his thirty-nine years, has an incredible résumé. He had been Facebook’s chief technology officer and was the engineer who came up with the iconic Like button. Prior to that, he’d been at Google, where he’d co-created Google Maps. Bret had also co-founded a start-up called Quip, which was an app that allowed real-time communication and collaboration on documents, lists, tasks, spreadsheets, and presentations, across many devices. When we bought Quip in 2016, we not only gained an amazing technologist and executive in Bret, but also a host of next-generation productivity tools that we could integrate into Salesforce’s suite of products.
Richard, too, embodies the spirit of a trailblazer. A world-renowned AI researcher, he has an impish smile and mop of carrot-colored hair and is sometimes affectionately referred to as “the boy scientist.” On many weekends, you’ll find this thirty-six-year-old German native loading his paraglider and jet surfboard into his station wagon before heading to the Bay. To see him rev up his motor and hurtle straight into a breaking wave, you’d never guess he was a leading AI expert on deep learning—that is, on teaching software to mimic the way the neurons in our brains retain and process information. He is no slouch in the field of natural-language processing either, with more than thirty-four thousand scholarly citations to prove it.
Few people on earth, even in Silicon Valley, would have understood a fraction of the bewildering acronyms and jargon Bret and Richard tossed around that night. These accomplished innovators—along with dozens of other brilliant new hires and longer-term employees—were an embodiment of our commitment to seeking innovation everywhere.
Thanks to their collective effort, we’ve been able to infuse AI into our products so seamlessly that customers barely even notice it’s there. Think about how effortless it is to use Amazon’s Alexa, or Apple’s Siri, or Google Assistant to check the weather or play music. Thanks to Qingqing Liu, our top mobile engineer, we now have something similar that can help people run their businesses. In just a few months, she took all the underlying AI technology, along with mind-blowing quantities of Salesforce data, and turned it into an amazing experience we now call Einstein Voice Assistant.
Now Einstein lives in the palm of my hand, where he functions as a voice-enabled digital assistant that understands the context of our business, attends our meetings, collects the relevant data in order to update notes and records related to the conversations, and even contributes comments. Our goal is that over time, we’ll be able to provide every customer with AI-powered digital assistants like these—and that they will grow more sophisticated as the technology evolves.
In coming years, more and more innovation will come from humans and machines working together, taking advantage of the unique capabilities of each. With machines doing more of the routine, repetitive work and the pattern recognition humans can’t, we’ll be free to spend more time trying to “be mindful and project the future” (as a wise man once advised me to do).
There are legitimate concerns about the ethics of AI, and I share them. But in the end, it’s humans who are crafting artificial intelligence, and these tools will be exactly as ethical as the people who build them. By the same token, machines don’t have a beginner’s mind, and no technology is either inherently good or inherently evil—what matters is how it’s used. The products of the future will only be as good as the honest, open conversations that happen around them.
The fears about AI and robotic automation eliminating jobs, too, are very real. This is why I believe that as computers take on more tasks previously performed by human workers, we need to find more ways to help people continuously adopt the curious, open-minded, adventurous spirit of a trailblazer, so they can learn and adapt to this brave new world ahead.
Like it or not, artificial intelligence is our future. And the only way it will work, at Salesforce or anywhere, is if the people designing the technology collaborate seamlessly with the people using it. Which makes it all the more critical that we fully embrace the fact that innovation truly does come from anywhere.
* * *
In 2017, just a few months after joining the company, Richard Socher approached me with a worried expression. As chief scientist, he didn’t understand how I expected him to do all the cutting-edge work we envisioned with the modest budget he’d been given.
“You’re not spending enough money on innovation,” he said with genuine puzzlement. “How can you call innovation a core value?”
At first I didn’t quite understand the question. In my view, the budget was fine. Then I remembered that Richard hadn’t been with us long enough to understand how innovation worked at Salesforce. He didn’t realize that our ability to build innovative products wasn’t solely a matter of how many dollars we threw at research and development. It was also about having a diverse ecosystem of curious people who challenged conventional wisdom and were empowered to pursue wild ideas.
Here’s how we do innovation at Salesforce, I told Richard: “We seek innovation everywhere.”
SIX
EQUALITY
A Good Look in the Mirror
In March 2015, when Salesforce’s employee success chief, Cindy Robbins, arrived at my home for one of the regular meetings I host with my senior executives, I could tell that something was a little off. Not only did she seem oddly reserved, even a bit anxious, she’d taken the unusual step of bringing backup, namely another senior woman executive, Leyla Seka.
I’ve never believed that the corporate office is the best backdrop for a relaxed, candid conversation, which is why I usually convene these one-on-one meetings at my home office. The standard attire is “casual,” which, in the vocabulary of tech companies, translates to “jeans.” On this day, however, I had taken “casual” to a new level. As it happened, I was competing in a charity Fitbit challenge against Dell’s founder and CEO Michael Dell and had waltzed into my meeting wearing shorts, a T-shirt, and a baseball cap. I could tell by the energy in the room that something important was clearly about to transpire, and I’d shown up looking like a gym rat.
Cindy an
d Leyla were both in their early forties and had grown up in the Bay Area. Cindy had joined Salesforce as a recruiter in 2006 and eventually rose to President and Chief People Officer, directly overseeing seven hundred employees. Leyla had signed on in 2008 as director of marketing for Salesforce AppExchange and gone on to a variety of roles in our business-software units.
Although they’d been close friends for years, they often laughed about how they hadn’t meshed at first. While Cindy is a self-described introvert with a reputation inside the company as a calm and steel-spined professional, Leyla, who is as extroverted as they come, wears Birkenstock sandals with designer scarves and once worked in the Peace Corps. After getting off to a rocky start, the pair had eventually bonded over the shared challenges of climbing the ladder in a male-dominated industry.
“What’s up?” I asked, tentatively.
If Cindy and Leyla were thrown off by my workout attire, they didn’t show it. They sat down and got right to the point. They’d come to tell me they suspected that women employees at Salesforce were being paid less than men for the same work.