Intel CIO Says AI Playing a Growing Role in Chipmaker's Operations
By Sara Castellanos
Intel Corp. Chief Information Officer Archana Deskus says
artificial intelligence has become more critical to the
semiconductor maker during the coronavirus pandemic.
AI is helping Intel generate insights and increase the speed at
which products are tested, which has been crucial as technology
initiatives were accelerated in recent months, Ms. Deskus said. The
Santa Clara, Calif.-based company has also benefited from
investments in data analytics and machine learning algorithms over
the past decade as it contends with the pandemic's effects on its
business, she said.
"We've spent a decade building up a sturdy data foundation which
is the basis for AI efforts, and when the pandemic hit, we were
able to move quickly," Ms. Deskus said.
Intel is also using AI algorithms to identify problems with
employees working remotely and for optimizing inventory throughout
its supply chain, as the pandemic caused manufacturing disruptions
in China, Ms. Deskus said.
By the end of 2024, 75% of organizations experimenting with AI
today are expected to shift from testing AI algorithms to fully
deploying them in business operations, according to a June report
from technology research firm Gartner Inc. That is partly because
AI provides companies with critical predictions that can drive
revenue growth and reduce cost and risk, according to Gartner.
AI has become more important to enterprise chief information
officers during the pandemic, and information technology executives
are deploying AI projects much faster than they would have in the
past, said Svetlana Sicular, vice president analyst at Gartner.
Some IT executives have recently deployed AI-based chatbots, for
example, to help call-center agents answer routine questions from
customers during the pandemic.
"Some projects were implemented unusually very quickly and they
will probably not return to those long cycles," she said. "Most
[CIOs] will see that AI is efficient where they need it."
Ms. Deskus became CIO of Intel on Jan. 30 and oversees its
global IT operations, including how technology is used in
manufacturing. Her team includes more than 5,000 employees. She was
previously CIO at Hewlett-Packard Enterprise and held CIO roles at
Baker Hughes, Timex and United Technologies.
A critical use case for AI in the coming months will be to speed
up product testing, Ms. Deskus said.
Intel's IT teams developed an AI-based hardware validation
program composed of more than 50 machine learning algorithms, an
effort that began about five years ago. The machine learning
algorithms can automatically detect hardware bugs quicker than
humans, because they are able to sift through terabytes of data to
find anomalies, Ms. Deskus said.
"We can do this in a much more sophisticated way than we could
previously," she said. Previously, the work of generating tests and
selecting which tests to execute was performed manually.
The algorithms are currently testing all of the functionalities
of Intel's future chips. Intel reported stronger earnings in its
most recent quarter but it has signaled a delay in its development
of superfast chips.
Earlier this year, Ms. Deskus oversaw the deployment of virtual
private networks to support approximately 100,000 employees who
began working remotely in early March. VPNs allow employees to work
on their computers securely from home. Intel's technology team used
machine-learning algorithms developed in-house as well as tools
built by a network monitoring company to proactively identify
issues with remote-work setups, she said.
The algorithms were used in part to detect problems with Wi-Fi
connectivity and internet traffic routing. Intel helped employees
optimize routing configurations and bandwidth use.
"We tried to help them get as similar of an experience as if
they were in the office," Ms. Deskus said.
Intel used AI algorithms to help simulate various supply chain
and logistics scenarios. For example, when supply chain disruptions
for Intel products occurred at factories in China during the height
of the pandemic, the company used machine-learning models to help
identify alternative supply chain routes.
AI algorithms were also used to identify the root cause of
quality issues on the manufacturing line in real-time to correct
those problems right away. Previously, Intel relied on data
analysis without machine learning to identify the root cause of
problems after manufacturing had been completed.
Before the pandemic, automation wasn't a major part of Intel's
manufacturing strategy, Ms. Deskus said. "The reliance on the
[automation] technology and the pivot is happening much faster than
if we didn't go through the crisis," she added.
Write to Sara Castellanos at email@example.com
(END) Dow Jones Newswires
September 01, 2020 18:03 ET (22:03 GMT)
Copyright (c) 2020 Dow Jones & Company, Inc.
Historical Stock Chart
From Sep 2020 to Oct 2020
Historical Stock Chart
From Oct 2019 to Oct 2020