Build 2017 takeaway - Microsoft transforms itself for a multi-devices, cross-platform, AI-powered future with Graph

After watching the Build 2017 keynotes and presentation, what surprised me the most, is the ambition and the potential of the Microsoft Graph. Seeing Microsoft Graph in action has answered quite a few questions at the back of my mind, and lessened my concern that Microsoft has apparently lifted its foot off the pedal on the mobile platform.

What I see in Microsoft Graph is the desire for Microsoft to maximize the value of its huge collection of data and services. It allows Microsoft to pool together everything it knows about an individual user across from various silos (all the services and apps that are ever offered by Microsoft: Office, Skype, Bing, LinkedIn, etc.) and "defrag" them into a coherent whole, and it allows Microsoft to transform all of its services from monolithic products into "microservices" - APIs that can be more flexibly recompose and accessed, both by itself, by users and by the 3rd party developers. Why is this necessary? A few guesses:

1. Data powers the economics and new wave of AI-driven experience. All the data put together is of greater value than sum of its parts. Data can be fed into AI to provide more valuable, previously unimaginable new services. There are many other ways to extract economic values, advertising being one of many. These data encompasses everything from user data, user's social connections to broader knowledge entities that Bing sourced from internet - locations, facts, news, etc. Graph could unify access to all of these data behind a common scheme.
2. In order for an AI assistant like Cortana to reach its full potential, it not only needs data from users, it also needs to perform tasks on behalf of users. For that, it needs API access to everything Microsoft products and services can do (Bing can return search results, Skype can send a chat, Office knows how to read a PowerPoint presentation and update calendars). Many of these have SDKs of various kinds today, but to scale, a unified, extensible APIs across everything makes more sense. Formalizing the API contracts into the graph is necessary for engineering but also opens the door for 3rd party developers add values down the line. This can transform the whole Microsoft from a set of monolithic products into huge collection of microservices behind Graph.
3. The world has moved to mobile, multi-devices and also multi-platforms. Every services and apps face the challenge of having to support to all the devices, platforms and modalities (think Echo devices that have no display). Moving as much functionalities into a shared code base in the cloud behind the graph simplifies the apps to mostly provide distinct UI code for Android, iOS, Windows and it also allows breaking off part of the app functionalities into various new micro-app paradigm like notification center applets, chat bots and conversational "skills" in Cortana / Echo. This work paves ways for the future when what constitute an app is being redefined.

As in any platform, you need a killer app to bring the strength, purpose and scale of the platform into focus. But at the dawn of the platform, you only need to start with a set of compelling apps and experience to attract a critical mass of users into this platform in time for the killer app to explode. As far as I can tell, that status of the killer app is Cortana's to lose, but honestly, it can be something else, and that won't take away from the potentials of the platform. With all the data and functionalities unified behind the Microsoft Graph, something awesome is going to happen one way or another.

From these lens, one can see the world as features that brings data into the Microsoft Graph, and features that project Microsoft Graph out to create user values. Naturally, all the Microsoft properties (LinkedIn included) will start funnel their data into the Microsoft Graph, and attracting 3rd party developers to do so as well will be important down the line. Microsoft Graph will be projected onto Windows, Android and iOS as deeply as possible. Windows being the platform that Microsoft owns, will be the showcase. Many of the new features showcased in Build are all projections of Graph: like OneDrive file on-demand, Timeline, cloud clipboard, etc. They aim to convert all Windows users into Graph users and then brings their Graph use to their other non-Windows devices like their iOS and Android phones. Anyone on Microsoft Graph would be a customer that Microsoft can provide user values in exchange of economics values down the line.

Questions I have been wondering:

• Why did Microsoft started making and acquiring a bunch of Android experience components? Like Arrow Launcher, SwiftKey keyboard to apps, etc.

I think these are the projection points for Microsoft Graph. The power of graph can be projected in wide range of user experience touch points, so casting a wide net is a good idea to set up these positions. From this point of view, keyboard, launchers are in the middle of various user scenario flows, and therefore are tremendously useful and attractive.

• Why didn't Microsoft try harder to save and push Windows Mobile?

I don't think one sets out to lose the mobile platform game if they can help it. But with the ambition of Microsoft Graph in sight, Microsoft Graph is just as powerful and lucrative as a platform for Microsoft to create and maintain as Windows if not more so.

Windows' economics as a desktop platform came from people paying for desktop OS to run apps they need, plus halo effect for first party apps like Office. Desktop apps moved to web and that has weakened the incentive for people to pay for desktop OS as long as they have a browser. Mobile OS's economics came from either selling hardware (iOS) or gathering user data to sell ads (Android), plus a cut of app store sales. Even if Microsoft has a viable market share with Windows Mobile, it's unclear whether it would have as strong economics as either iOS or Android. I think not having a strong horse in the mobile OS race allowed Microsoft pivot into the cross-devices, cross-platforms Graph, which could provide superior economics once usage and data volume per user reach critical mass.

In conclusion - I think Microsoft Graph is the story that I am the most excited about from Build 2017 for the consumer business. I think Microsoft has a new platform that transcends Windows in its infancy and its economic potential may be able to offset the decline of desktop PC OS revenue. Given the number of sockets for Microsoft Graph is inherently larger than Windows user base, it may even surpass Windows one day in terms of profitability.