Quick Thoughts: Alexa and Cortana Sitting in a Tree…

Print Friendly, PDF & Email


Paul Sagawa / Tejas Raut Dessai

203.901.1633 /203.553.9827

psagawa@ / trdessai@ssrllc.com


September 5, 2017

Quick Thoughts: Alexa and Cortana Sitting in a Tree …

  • AMZN and MSFT will link their AI assistants, a “chocolate and peanut butter” combo that improves the reach and capabilities of both platforms
  • AMZN can sell Alexa into MSFT’s untapped base of 500M Windows 10 users and can offer Office functionality (e.g. email, calendar, etc.) to Echo users via Cortana
  • MSFT can kick-start Cortana for consumers, both as a skill on Alexa and with a more complete solution for PC users, complementing its push to establish it as a standard in enterprise computing.
  • For both, the partnership should drive greater and more diverse query volume, yielding data to improve their AI assistants in the face of competition from GOOGL and AAPL

Jeff Bezos sidled up to Satya Nadella at the May 2016 Microsoft CEO Summit with an idea. Each company had an AI virtual assistant product at the core of its strategy, but the two efforts were aimed at very different end markets and each had strength in a different set of core services. Why not partner to reach a larger audience with a full set of services?

Amazon’s Alexa was entirely consumer focused, a pragmatic voice interface for the company’s increasingly popular Echo home speakers, linked to a growing set of pre-programmed skills provided by Amazon and its ecosystem of 3rd party partners. First mover advantage, the powerful Amazon retail machine, and those useful skills was keeping the company ahead of rivals Google and Apple, but those companies had their own strengths. Google’s Android and Apple’s iOS gave each of them installed bases of more than a billion potential users. Siri had been standard on iPhones and iPads since 2014 and had a substantial share of mind. Google was late to the party with Assistant, but had extraordinary assets – a data base of trillions of typed search queries, access to detailed personal information on its users, and the deepest roster of AI talent in the world. Despite Amazon’s own considerable AI capabilities, Bezos was concerned that the dominant smartphone platforms could freeze him out with universe of consumers without Echo home speakers, while catching up on building an ecosystem of service partners, and even surpassing Alexa on interpreting natural language commands and anticipating needs.

Microsoft could help. Nadella had pivoted away from the consumer focus of his predecessor, dropping its smartphone OS ambitions and focusing on leading a cloud/AI revolution in the enterprise market. Cortana was a big part of that – eventually Microsoft planned to make it a blanket user interface for Office and all its other application software. There were well over a billion Office users, many running the application at home as well as at work. The problem? Enterprises move very slowly – converting those billion Office users to Cortana users running Windows 10 on PCs with microphones, then getting users to exploit the benefits that an AI assistant like Cortana could provide will take a long time. Meanwhile, consumer markets move fast – Amazon has already sold over 20 million Echos, while Google and Apple have those huge installed bases. Cortana also lacked for consumer services outside of its Bing search engine (which already powers Alexa’s search) and controls for its Xbox One video game console. Microsoft had a nice party dress but no invitation to the ball.

The fruit of the partnership begun at that conclave a year ago is now beginning to ripen. Alexa users will be able to call up Cortana to check on their emails, their schedule, and find Office files. Eventually, they will be able to control Microsoft’s growing roster of SaaS applications, including LinkedIn. Cortana users – yes, there are Cortana users – will be able to use their PCs and Xboxes to control the lights, call Uber cars, and order stuff from Amazon via Alexa. The possible reach for each platform grows and each platform can offer users a richer spate of services – all without very much competitive overlap. At the same time, both companies will likely field more queries on a wider array of topics, all the better to push the natural language interpretation AIs forward.

This is good for both Amazon and Microsoft and slightly bad for both Google and Apple. It is very early in the AI virtual assistant wars, but this partnership makes for a more formidable competitive environment. I expect Google to fire the next shot – it needs to lock in a lot more content and services partners for its limply named Google Assistant, or its superior AI-driven speech recognition, query interpretation, and request anticipation could be for naught. Still, Assistant is an obvious priority for Google, which has a lot of firepower to bring to bear on the market.

Meanwhile, Apple doesn’t seem to have made its assistant much of a priority – it introduced Siri in 2011 and has frittered away any first mover advantage. Apple’s World-Wide Developer’s Conference has largely back-burnered Siri for the iPhone and Apple Watch. The expensive HomePod announced in the Spring is focused on sound quality rather than personal assistance, competing more directly with Sonos speakers than with the Amazon Echo. Apple’s tight privacy policy restricts its use of user queries in making Siri better or giving it reach across devices. Rather than burn its AI resources on self-driving car software, or its attention on chasing Hollywood, I think Apple might be better served by doubling down on its investment in Siri while the game is still afoot.


Print Friendly, PDF & Email