|Alphabet and its best-known subsidiary, Google, have put considerable resources into machine learning going back to 1999, the first year that Google acknowledged publicly that it used AI to improve Google Search, then its only product. Once Google decided to get more serious about its cloud computing business and serving enterprise customers—Google Cloud storage officially launched in 2010—it has found more ways to take its AI investment and acumen and use it to serve others.|
How Amazon, Google, Microsoft, And IBM Sell AI As A Service
The tech giants with cloud computing businesses are using artificial intelligence offerings to distinguish themselves and win business.
By Fast Company Staff
The success of Amazon’s Alexa voice assistant has reverberated throughout the business world, making AI- powered chat the next big thing. [Illustration: Daniel Zender]
Alphabet, Amazon, and Microsoft have all discovered that the artificial intelligence they use to make their own products better can be turned into a service and sold to corporate customers as a value-added service on top of their booming cloud-computing businesses.
Alphabet and its best-known subsidiary, Google, have put considerable resources into machine learning going back to 1999, the first year that Google acknowledged publicly that it used AI to improve Google Search, then its only product. Once Google decided to get more serious about its cloud computing business and serving enterprise customers—Google Cloud storage officially launched in 2010—it has found more ways to take its AI investment and acumen and use it to serve others. Diane Greene, SVP of Google Cloud, has admitted that enterprise customers had been wary of Google because the company has been so consumer focused; its AI capabilities have played a meaningful role in winning them over.
Alphabet has two major divisions working on AI: Google Brain and DeepMind, which it acquired for $500 million in 2014. Both groups have worked on applying AI in healthcare, for example, which then allows Google Cloud to better serve businesses in that field. The company’s efforts in image recognition can become valuable for Airbus and other aerospace businesses that need to process and glean insights from large volumes of satellite imagery. All of Google’s work on Google Translate can now help any global business with a call center. Although most of the value in Google’s AI accrues to its own products and services, the company has stated that Google Cloud is one of its fastest-growing business units.
Amazon has a much more natural synergy between its AI efforts and how it can sell those initiatives to others via its industry-leading cloud computing service. As CEO Jeff Bezos wrote earlier this year in his letter to shareholders, “Much of what we do with machine learning happens beneath the surface . . . quietly but meaningfully improving core operations.” The examples Bezos cites include demand forecasting, fraud detection, and translations—all features that any business would value. As our feature on the Great AI War recounts, a sheriff’s department in Oregon pays Amazon about $6 a month to use Amazon’s facial-recognition service on an ongoing basis.
More than any of its rivals, Amazon has electrified the public with its audacious vision for an AI-powered future. Its line of Echo devices, brought to life by the artificially intelligent Alexa, has defined the path for the next generation of home automation and commerce and made voice-powered speakers arguably the hottest segment in consumer electronics. That success has enabled Amazon to release the technology powering Alexa as its own product so that any company can develop its own intelligent voice applications.
This strategy is central to Amazon’s history of success; it has largely always relied upon its ability to transform something it built for itself into something it can then sell to millions of businesses. Amazon started as a mere bookseller and then opened up its marketplace to let other retailers take advantage of its e-commerce platform. After it built warehouses to fulfill orders for customers, it offered Fulfillment by Amazon to those same marketplace businesses. Amazon Web Services started because Amazon had had to build excess computing capacity to support its business during the busiest shopping season; it could then sell that capacity to a host of others. This is how Amazon’s famous “flywheel” works and AI-powered services are its next frontier.
To that end, keep a close eye on the company’s retail concept called Go. It relies on computer vision and machine learning to present a different kind of shopping experience. Amazon has yet to open this new take on the convenience store to the public almost a year after announcing the idea. But once the company gets Go working, do not expect the company to roll out thousands of Go stores across the country. It is far more likely that Amazon will offer up this AI-powered retail infrastructure to existing shopkeepers who will pay Amazon a recurring fee to use it.
Also note that Amazon Web Services currently represents almost 10% of the company’s annual revenue and it is a part of Amazon’s business that investors monitor very closely. The more Amazon can keep AWS humming, the more its entire enterprise thrives.
Unlike Alphabet/Google or Amazon, almost all of Microsoft’s business lies in serving enterprise customers. It is the tech giant most focused on converting AI directly into revenue. “Our company’s identity is fundamentally about creating technology so that others can create more technology,” CEO Satya Nadella told Fast Company recently. “And it’s essential that it is being used for empowering more people.”
Artificial intelligence “is at the intersection of our ambitions,” Nadella told an audience of Microsoft partners in September 2016, suggesting that it will let the company “reason over large amounts of data and convert that into intelligence.” A few months later, Microsoft officially closed its $26.2 billion acquisition of LinkedIn, giving the company a large amount of data about employees, companies, and recruiting to reason over and try to make smarter.
In August, it debuted a real-time AI system for its enterprise cloud customers, which could help the company win business from companies who want to deploy such business initiatives as dynamic pricing and retail personalization. Microsoft’s mission to help companies in a wide range of industries to be more productive and effective means that it is the one company whose AI work is most keenly connected to its future prospects.
Similarly, IBM’s approach has been to target specific industries, from healthcare to retail, and learn those domains so that its Watson-branded AI (which IBM calls cognitive computing) can alleviate drudge work and wrangle impossibly large sets of data. “There’s a reason we call it cognitive [computing],” IBM CEO Ginni Rometty told the CNBC personality Jim Cramer in June 2017. “It’s about augmenting what you and I do so we can do what we’re supposed to, our best.”
IBM’s argument to customers is that it is the only company offering sector-based AI solutions and those businesses within them can own their own AI rather than just rent it. It’s also made the most overt effort to connect its industrial internet of things initiative to Watson, as best seen in IBM’s 2016 acquisition of The Weather Company for approximately $2 billion. The deal gave IBM access to 2.2 billion forecast points worldwide, a trove of data that Watson churns through to fuel multiple client services. These efforts have generated a lot of attention and Watson is arguably the strongest brand in AI, but they haven’t yet turned around IBM’s business.
A version of this article appeared in the November 2017 issue of Fast Company magazine.