LinkedIn could supply data to help Office better understand us
through our contacts
By Jay Greene
Microsoft Corp.'s $26.2 billion deal to buy LinkedIn Corp. is
the latest turn in a race by the software giant and its competitors
to mine vast quantities of data -- searches, emails, social
networks, browsing behavior -- for insights that would help their
products serve and even anticipate customer needs.
Microsoft, Apple Inc. and Alphabet Inc. are moving deeper into
using machine learning, in which computers rapidly examine existing
information in search of patterns that their products can use to
understand new information as it arrives.
As Microsoft unveiled its acquisition, Apple was showing off a
plan to bring advanced computer vision to its photo software. The
idea: let an iPhone identify family and friends to help categorize
images more easily. Last month, Alphabet unveiled Google assistant,
a software helper designed to predict what users want by analyzing
their behavior. It would learn their preferences so that it might
offer to make dinner reservations when a restaurant was mentioned
in a text message.
With LinkedIn, Microsoft gains a trove of details about the work
lives, colleagues, and employers of the business social network's
105.5 million monthly active users: a unique set of data that the
software giant hopes will give it an advantage in serving business
customers.
"That's really what this next wave of technology innovation is
about," Microsoft Chief Executive Satya Nadella said during a
conference call on Monday, referring to machine learning. "But in
order to be able to do that, you need data, and LinkedIn represents
that."
Microsoft, which has embraced the spread of "intelligence"
throughout its offerings as an explicit goal, aims to enhance,
rather than replace, human performance. Where Google's autonomous
cars stop, turn, and change routes without human intervention,
Microsoft has focused on using machine intelligence to provide
insights that can inform human decisions.
Forrester Research Inc. analyst Jeffrey Hammond refers to
Microsoft's approach as "augmented intelligence" to distinguish it
from artificial intelligence. "It adds a whole lot of rich
contextual information," Mr. Hammond said.
The fruits of this approach are already available in the
Redmond, Wash., company's products. Last month, it revamped its
SharePoint collaboration and networking program to analyze lists of
meeting attendees, frequent email correspondents, collaborators on
documents, and the like. The software uses the results to deliver
information, such as internal reports or news articles, that might
be useful to a given user.
If Microsoft can deliver on that promise, LinkedIn data will
feed Microsoft products such as SharePoint, the Office productivity
suite, or Dynamics sales tools, to streamline projects and pinpoint
prospects. Microsoft, as part of its presentation to analysts about
the deal, described a scenario in which Cortana, its digital
assistant, notified a user that a coming meeting included someone
who attended the same college and shared a LinkedIn connection.
Then, Cortana offered to link to the attendee's profile and share
the meeting presentation.
Other tech giants have their own versions of Microsoft's Cortana
strategy. Apple's voice-activated assistant Siri is now smart
enough to work with non-Apple apps to perform tasks like calling
for car service or sending messages through text networks like
WeChat. Facebook is fine-tuning a concierge called M whose digital
brain passes along to human attendants any requests that it isn't
sure it can fulfill. IBM's Watson has been studying medical data in
hope of helping doctors make faster, more accurate diagnoses.
Amazon.com Inc.'s Echo, a smart household gadget in the form of
a small cylinder, functions as ears and voice for Alexa, an
artificial intelligence that answers spoken questions about topics
like weather and traffic. Google intends to roll out Google Home,
an Echo competitor that taps the search giant's vast data
repository to understand what users want.
Like Microsoft, all the tech behemoths aim to distribute
intelligence throughout their products.
"The next big step will be for the very concept of the 'device'
to fade away. Over time, the computer itself -- whatever its form
factor -- will be an intelligent assistant helping you through your
day," Google CEO Sundar Pichai wrote in an April letter to
shareholders.
Artificial intelligence, though, has proven remarkably foolish
at times. In March, Microsoft introduced Tay, a Twitter feed that
mimicked the texting habits of an American woman aged 18 to 24. But
online pranksters gamed Tay's learning mechanism, which led the bot
to tweet anti-Semitic rants.
As tech giants train their machine learning engines on ever
larger stores of data, the potential grows for users to feel
unnerved. To derive value from LinkedIn, Microsoft will need to
keep tabs on users' contacts, communications and activities. Its AI
will be informed by updates to LinkedIn profiles, which may
indicate that a user is considering a career move, and project
management systems, which may contain designs of future products
and other strategically sensitive information.
"How do they use that data to provide a truly convenient
experience without crossing the border into creepy ones?" asks
Forrester's Mr. Hammond.
At LinkedIn, users control how they share their data. Mr.
Nadella said that customers would decide what to share. But
Microsoft, like its peers that occupy the center of the digital
universe, is betting that those customers will let it mine their
information because they'll get value from the insights it
generates.
"When it comes to machine learning and analytics, it is a
tremendous opportunity for us," Mr. Nadella said.
Deepa Seetharaman, Jack Nicas and Daisuke Wakabayashi
contributed to this article.
Write to Jay Greene at Jay.Greene@wsj.com
(END) Dow Jones Newswires
June 15, 2016 02:49 ET (06:49 GMT)
Copyright (c) 2016 Dow Jones & Company, Inc.
Linkedin Corp. Class A (NYSE:LNKD)
Historical Stock Chart
From Jun 2024 to Jul 2024
Linkedin Corp. Class A (NYSE:LNKD)
Historical Stock Chart
From Jul 2023 to Jul 2024