What's Now and Next in Analytics, AI, and Automation

9/3/2017 Whatsnowandnextinanalytics,AI,andautomation|McKinsey&Company
ExecutiveBriefing
McKinseyGlobalInstitute
May2017
Whats now and next in analytics, AI,

and automation
Innovations in digitization, analytics, artificial intelligence, and

automation are creating performance and productivity
opportunities for business and the economy, even as they reshape
employment and the future of work.
R
apid technological advances in digitization and data and analytics have been
reshaping the business landscape, supercharging performance, and enabling the
emergence of new business innovations and new forms of competition. At the same time,
the technology itself continues to evolve, bringing new waves of advances in robotics,
analytics, and artificial intelligence (AI), and especially machine learning. Together they
amount to a step change in technical capabilities that could have profound implications
for business, for the economy, and more broadly, for society.
Table of contents
1. The opportunity available now

You are now signed in
2. The next wave of opportunity
Feel free to select your topics of interest or change your
password at any time.
3. What about employment and work?
http://www.mckinsey.com/globalthemes/digitaldisruption/whatsnowandnextinanalyticsaiandautomation 1/21
4. What should leaders do?
The opportunity available now

Some companies are gaining a competitive edge with their use of data and analytics,
which can enable faster and larger-scale evidence-based decision making, insight
generation, and process optimization. But there is room to catch up and to excel.
Harnessing digitizations potential is similarly uneven.
Data and analytics are transformational, yet many companies

are capturing only a fraction of their value
Data and analytics have been changing the basis of competition in the years since our first
report on big data in 2011. Leading companies are using their capabilities not only to
improve their core operations but also to launch entirely new business models. The
network effects of digital platforms are creating a winner-take-most dynamic in some
markets. Yet while the volume of available data has grown exponentially in recent years,
most companies are capturing only a fraction of the potential value in terms of revenue
and profit gains.
Effective data and analytics transformations have several components:

Asking fundamental questions to shape the strategic vision: What will
data and analytics
Feel be
freeused for?your
to select Howtopics
will the insights
of interest ordrive value?
change your
password
Which data sets are most useful for the at any time.
insights needed?
Solving for the problems in the way data is generated, collected, and
organized. Many incumbents struggle to switch from legacy data
systems to a more nimble and flexible architecture that can get the
most out of big data and analytics. They may also need to digitize
their operations more fully in order to capture more data from their
customer interactions, supply chains, equipment, and internal
processes.
Acquiring the skills needed to derive insights from data;

organizations may choose to add in-house capabilities or outsource
to specialists.
Changing business processes to incorporate data insights into the

actual workflow. This is a common stumbling block. It requires
getting the right data insights into the hands of decision makersand
making sure that these executives and mid-level managers
understand how to use data-driven insights.
Putting all these components in place is not easy. In a recent McKinsey survey of more
than 500 executives representing companies across the spectrum of industries, regions,
and sizes, more than 85% acknowledged that they were only somewhat effective at
meeting goals they set for their data and analytics initiatives.
Data and analytics are disrupting business models and bringing

performance benefits
Disruptive data-driven models and capabilities are reshaping some industries, and could
transform many more. Certain characteristics of a given market open the door to
disruption by those using new data-driven approaches, including:

inefficient matching of supply and demand
prevalence of underutilized assets
dependence on large amounts of demographic data when behavioral

data is now available
human biases and errors in a data-rich environment
In industries where most incumbents have become used to relying on a certain kind of
standardized data to make decisions, bringing in fresh types of data sets (orthogonal
data) to supplement those already in use can change the basis of competition. We see
this playing out for example in property and casualty insurance, where new companies
have entered the marketplace with telematics data that provides insight into driving
behavior, beyond the demographic data that had previously been used for underwriting.
One of the most powerful uses is micro-segmentation based on behavioral characteristics

of individuals. This is changing the fundamentals of competition in many sectors,
including education, travel and leisure, media, retail, and advertising.
Digitization, more broadly, is also progressing unevenly among

companies, sectors, and economies
The corporate worlds broader embrace of digitization is similarly uneven. Our use of the
term digitization (and our measurement of it), encompasses:
Assets, including infrastructure, connected machines, data, and data

platforms, etc.,
Operations, including processes, payments and business models,

customer and supply chain interactions and
The workforce, including worker use of digital tools, digitally-skilled

workers, new digital jobs, and roles. In measuring each of these
You are now

even among big companies (Exhibit 1). signed in
various aspects of digitization, we find relatively large disparities

Exhibit1
Our research finds that companies with advanced digital capabilities across assets,
operations, and workforces grow revenue and market shares faster than peers. They

improve profit margins three times more rapidly than average and, more often than not,
have been the fastest innovators and the disruptors in their sectorsand in some cases
beyond them.
Many of these top performers were born digital, but perhaps more impressive are the
smaller set of incumbent companies that have actively transformed themselves into
digital leaders and benefit doubly from their traditional strengths and their new digital
capabilities.
There are also disparities between sectors in terms of degree of digitization:
In the United States, the information and communications

technology (ICT) sector, media, financial services, and professional
services are surging ahead, while utilities, mining, and
manufacturing, among others, are in the early stages of digitizing. In
labor-intensive industries such as retail and health care, substantial
parts of their large workforces do not use technology extensively.
This unevenness can also be observed across countries; all have

significant room to increase their digitization:
The US economy as a whole is reaching only 18% of its digital

potential;
France has achieved 12% of its digital potential, the European

Union average, while Germany and Italy are at 10%;
Emerging economies are even further behind, with countries

in the Middle East and Brazil capturing less than 10% of their
digital potential.
Digitization is transforming globalization, creating

opportunities now for companies and economies

The world is more connected than ever, but the nature of its connections has changed in a
fundamental way. The amount of cross-border data flows has grown 45 times larger since
Feel
just 2005. It is projected to free toby
increase select your topics
an additional oftimes
nine interest
overor change
the your
next five years as
flows of information, searches, communication, video, transactions, and intracompany

traffic continue to surge.
In addition to transmitting valuable streams of information and ideas in their own right,
data flows enable the movement of goods, services, finance, and people. Virtually every
type of cross-border transaction now has a digital component.
Approximately 12% of the global goods trade is conducted via international e-commerce,
with much of it driven by platforms such as Alibaba, Amazon, eBay, Flipkart, and
Rakuten. Beyond e-commerce, digital platforms for both traditional employment and
freelance assignments are beginning to create a more global labor market. Some 50% of
the worlds traded services are already digitized. These transformations enable small and
medium-sized enterprises around the world to compete head to head with larger
industry incumbents.
Back to top
The next wave of opportunity

Coming over the horizon is a new wave of opportunity related to the use of robotics,
machine learning, and AI. Companies that deploy automation technologies can realize
substantial performance gains and take the lead in their industries, even as their efforts

contribute to economy-level increases in productivity.
Advances in robotics, AI,toand

Feel free machine
select learning
your topics herald
of interest a new
or change your era
of breakthrough innovation andpassword
opportunity
at any time.
Recent advances in robotics, machine learning, and AI are pushing the frontier of what
machines are capable of doing in all facets of business and the economy.
Physical robots have been around for a long time in manufacturing, but more capable,
more flexible, safer, and less expensive robots are now engaging in ever expanding
activities and combining both mechanization, cognitive and learning capabilitiesand
improving over time as they are trained by their human coworkers on the shop floor, or
increasingly learn by themselves.
The idea of AI is not new, but the pace of recent breakthroughs is. Three factors are
driving this acceleration:
Machine-learning algorithms have progressed in recent years,

especially through the development of deep learning and
reinforcement-learning techniques based on neural networks.
Computing capacity has become available to train larger and more

complex models much faster. Graphics processing units (GPUs),
originally designed to render the computer graphics in video games,
have been repurposed to execute the data and algorithm crunching
required for machine learning at speeds many times faster than
traditional processor chips. More silicon-level advances beyond the
current generation of GPUs are already emerging, such as Tensor
Units. This compute capacity has been aggregated in hyper-scalable
data centers and is accessible to users through the cloud.
Massive amounts of data that can be used to train machine learning

models are being generated, for example through daily creation of
billions of images, online click streams, voice and video, mobile

locations, and sensors embedded in the Internet of Things.
The combination of these breakthroughs has led to spectacular demonstrations like

Feeldefeated
DeepMinds AlphaGo, which free to select your
a human topics ofof
champion interest or change
the complex boardyour
game Go
in March 2016. Googles DeepMind and the password
Universityat
ofany time.
Oxford applied deep learning to
a huge data set of BBC programs in 2016 to create a lip-reading system that is more
accurate than a professional lip reader.
Formidable technological challenges must still be overcome before machines can match
human performance across the range of cognitive activities. One of the biggest technical
challenges is for machines to acquire the capability to understand and generate natural
languagecapabilities that are indispensable for a multitude of work activities. Digital
personal assistants such as Apples Siri, Amazons Alexa, and Google Assistant, are still in
developmentand often imperfecteven though their progress is palpable for millions
of smartphone users.
Harnessing these evolving technologies will unlock multiple

benefits for companies
For companies, successful adoption of these evolving technologies will significantly
enhance performance. Some of the gains will come from labor substitution, but
automation also has the potential to enhance productivity, raise throughput, improve
predictions, outcomes, accuracy, and optimization, as well expand the discovery of new
solutions in massively complex areas such as synthetic biology and material science.
Already today, a range of automation technologies is generating real value. For example:
Rio Tinto has deployed automated haul trucks and drilling machines
at its mines in Pilbara, Australia, and says it is seeing 1020%
increases in utilization there.
Google has applied artificial intelligence from its DeepMind machine

learning to its own data centers, cutting the amount of energy they
use by 40%.

In financial services, automation in the form of straight-through
processing, where transaction workflows are digitized end-to-end,
can increase the scalability of transaction throughput by 80%, while
reducing errors by half.
Furthermore, a plethora of machine learning business use cases are emerging across
sectors (Exhibit 2).
Exhibit2

Scenarios we developed for several settings, including a hospital emergency department,
aircraft maintenance, oil and gas operations, a grocery store, and mortgage brokering,
show that the value of the potential benefits of automationcalculated as a percentage of
operating costscould range from between 1015% for a hospital emergency department
to 25% for aircraft maintenance, and to more than 90% for mortgage origination.
AI and Automation will provide a much-needed boost to global

productivity and may help some moonshot challenges
The application of AI and the automation of activities can enable productivity growth and
other benefits not just for businesses, but also for entire economies. At a macroeconomic
level, based on our scenario modeling, we estimate automation alone could raise
productivity growth on a global basis by 0.8% to 1.4% annually.
AI and other technologies can also be broadly beneficial for society by helping tackle
some moonshot challenges, including climate change or curing disease. AI is already
being deployed in synthetic biology, cancer research, climate science, and material
science. For example, researchers at McMaster and Vanderbilt University have used
computers to exceed the human standard in predicting the most effective treatment for
major depressive disorders and eventual outcomes of breast cancer patients.
Back to top
What about employment and work?

The advent of a new automation age is raising public concerns about the effect on
employment and the future of work. For most occupations, partial automation is more
likely than full automation in free
Feel the medium
to selectterm,
yourand theof
topics technologies
interest orwill provide
change yournew
opportunities for job creation. password at any time.
About half the activities carried out by workers today have the
About half the activities carried out by workers today have the
potential to be automated
To assess the employment implications of automation, we focused on work activities
rather than whole occupations as a starting point. We consider work activities to be a
useful measure since occupations are aggregations of different activities, where each
discrete activity has a different potential for automation. For example, a retail
salesperson will spend some time interacting with customers, stocking shelves, or ringing
up sales.

Activities that are more easily automatable include physical activities in highly
predictable and structured environments, as well as data collection and data processing
(Exhibit 3). These activities exist across the entire spectrum of sectors, as this data
visualization of the automation potential of individual sectors shows.
Exhibit3

Our analysis of the automation potential extends to 46 countries representing about 80%
of the global workforce. Overall, we estimate that about half of the activities that people
are paid almost $15 trillion to do in the global economy have the potential to be
automated by adapting currently demonstrated technology. This data visualization of
global automation potential shows sizable differences between countries, based mainly
on the structure of their economies, the relative level of wages, and the size and dynamics
of the workforce.
All occupations will be affected. Only a small proportion of all occupations, about 5%,
consist of 100% of activities that are fully automatable using currently demonstrated
technologies. However, we find that about 30% of the activities in 60% of all occupations
could be automated (Exhibit 4). This means that many workers will work alongside
rapidly evolving machines, which will require worker skills also evolve. This rapid
evolution in the nature of work will affect everyone from welders to landscape gardeners,
mortgage brokersand CEOs; we estimate about 25% of CEOs time is currently spent on
activities that machines could do, such as analyzing reports and data to inform decisions.
Several key factors will influence the pace and extent of automation. These include:
Technical feasibility of automation, a critical first step that will

depend on sustained breakthrough innovation, but alone is not
sufficient;
Cost of developing and deploying solutions;
Labor market dynamics, including supply and demand, and costs of

human labor as an alternative to automation;
Business and economic benefits, not merely labor substitution

benefits but also benefits from new capabilities that go beyond
human capabilities;

Regulatory, user and social acceptance, which can affect the rate of
adoption even when deployment
Feel free makes
to select your topicsbusiness and
of interest economic
or change your
sense. password at any time.
A useful analogy to consider is that electric vehicles were demonstrated to be technically

feasible several decades ago, but it was not until some of these other factors became
realistic that they showed up on the road.

Exhibit4

Technology will also help create new jobs and new opportunities
Technology will also help create new jobs and new opportunities
for generating income, and will help labor markets function
better
The scale of shifts in the labor force over many decades that automation technologies will
likely unleash is of a similar order of magnitude to the long-term technology-enabled
shifts in the developed countries workforces as they moved most workers from farms to
factories and service jobs. Those shifts did not result in long-term mass unemployment
because they were accompanied by the creation of new types of work not foreseen at the
time. We cannot definitively say whether historical precedent will be repeated this time.
But our analysis shows that humans will still be needed in the workforce.
So even while technologies replace some jobs, they are creating new work in industries
that most of us cannot even imagine, as well as new ways to generate income and match
talent to jobs.
One third of new jobs created in the United States in the past 25 years were types that did
not previously exist, or barely existed, in areas including IT development, hardware
manufacturing, app creation, and IT systems management. The growing role of big data
in the economy and business will create a significant need for statisticians and data
analysts, for example; we estimate a shortfall of up to 250,000 data scientists in the US in
a decade.
Technology helps work in other ways. Digital talent platforms such as LinkedIn have
already begun to improve matching of workers with jobs, creating transparency and
efficiency in labor markets, and thereby raising GDP. While it is early days, there is
already evidence that such platforms can raise labor participation and working hours.
While independent work is nothing new (and self-employment is still the predominant
form of work in emerging economies), the digital enablement of it is. Our research finds
that 20% to 30% of the working age population in the US and the European Union is

engaged in independent work. Just over half of these workers supplement their income
and have traditional jobs, or are students, retirees, or caregivers. While 70% choose this
type of work, 30% turn to it out of necessity because they cannot find a traditional job at
Feel
all, or one that meets their free to
income select
and your topics
flexibility needs. of
Theinterest or change
proportion your
of independent
work that is conducted on digital platforms, while only about 15% of independent work
overall, is growing rapidly, driven by the scale, efficiency, and ease of use for workers and
customers that these platforms enable.
Those who pursue independent work (digitally enabled or not) out of preference are
generally satisfied, although those who pursue it out of necessity are unsatisfied with the
income variability and the lack of benefits typically associated with traditional work.
Back to top
What should leaders do?

Business leaders and policy makers have an imperative to find ways to harness the
potential of these technologies, even as they will have to address the significant
challenges.
Business leaders
For businesses, the opportunities are clear. Leaders should embrace the transformation
and performance opportunities already available to them (and their competitors) from
data, analytics, and digitization, as well as the rapidly evolving opportunities in AI,

robotics, and automation. To harness these benefits, business leaders will not only have
to invest in technology, but also in transforming their organizations. Specific approaches
will vary business by business, however several new mindsets will be critical:
Testing, experimenting, learning, and scaling fast: Beyond book

knowledge, business leaders will need to amass practical knowledge
from devoting resources to experiments applying technologies to real
problems, and then scaling those that show promise.
Reimagining business models and business processes: To make

full use of the power of analytics, AI, and other digital technologies
will require a thorough reimagining of processes, with priorities for
which processes to transform. Similarly, leaders will need to
reimagine how current business models could be transformed and
how new business models could be created based on these
capabilities.
Digital assets and capabilities as the new balance sheet: These

assets and capabilities, both hard and soft, are increasingly becoming
a competitive differentiator and platforms for innovation and
disruption. Each business regardless of industry and sector will likely
need to assess how distinctive its digital assets and capabilities are vs.
those of competitors.
Staying calibrated and investing accordingly: When it comes to

digital capabilities and progress on digitization initiatives, all too
often business leaders are satisfied with progress vs. their own past.
The most relevant calibration will be relative to the scale of the
opportunity and vs. competitors and potential disruptors both from
within their sectors and from outside them.

A new focus on human capital, including integrating workers and
machines: Companies are likely to face gaps in skills they need in a
more technology-enabled workplace, and would benefit from playing
a more active role in education and training. Humans and machines

will need to work together much more closely. That will require
retraining and often redeploying workers.
Policy makers and business leaders concerned with wider

economic and societal implications
Policy makers also have a powerful incentive to embrace the productivity growth
opportunity for their economies that these technologies offer. This will help ensure
future prosperity, and create the surpluses that can be used to assist workers and society
adapt to these rapid changes. At the same time, policy makers must evolve and innovate
policies that help workers and institutions adapt to the impact on employment:
Adopting policies to encourage investment: Through tax benefits

and other incentives, policy makers can encourage companies to
invest in human capital. Policy makers could accelerate the creation
of jobs in general through stimulating investment, and accelerate
creation of digital jobs in particular.
Encouraging new forms of entrepreneurship and more rapid new

business formation: Digitally enabled opportunities for individuals
to earn incomes. In addition, accelerating the rate of new business
formation will be critical. This will likely require simplifying
regulations, creating tax and other incentives.
Publicprivate partnerships to stimulate infrastructure

investment: The lack of enabling digital infrastructure is holding
back the digital benefits for some emerging economiesand even
are now signed in

underserved regions in developed countries. Publicprivate
You
partnerships could help address market failures.

Rethinking education, training, and learning: Policy makers

working with education providers could do more to improve basic
science, technology, engineering, and math (STEM) skills through the
school systems, and put a new emphasis on creativity as well as
critical and systems thinking.
Rethinking income support and safety nets: If automation (full or

partial) does result in a significant reduction in employment and/or
greater pressure on wages, some ideas such as universal basic income,
conditional transfers, and adapted social safety nets may need to be
considered and tested.
Incent investment in human capital: A broad range of incentives

exist for businesses to make capital and R&D incentives. Something
similar needed to encourage investment in human capital.
Back to top
A full version of this briefing note is available as a PDF download.
About the author(s)
James Manyika is director of the McKinsey Global Institute and a senior

partner at McKinsey & Company, based in San Francisco. MGI partners
Michael Chui, Anu Madgavkar, and Susan Lund contributed to this

briefing note.


What's Now and Next in Analytics, AI, and Automation

Uploaded by

Copyright:

Available Formats

What's Now and Next in Analytics, AI, and Automation

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

What's Now and Next in Analytics, AI, and Automation

Uploaded by

Copyright:

Available Formats

9/3/2017 Whatsnowandnextinanalytics,AI,andautomation|McKinsey&Company

Whats now and next in analytics, AI,

Innovations in digitization, analytics, artificial intelligence, and

1. The opportunity available now

4. What should leaders do?

The opportunity available now

Data and analytics are transformational, yet many companies

Effective data and analytics transformations have several components:

You are now signed in

Acquiring the skills needed to derive insights from data;

Changing business processes to incorporate data insights into the

Data and analytics are disrupting business models and bringing

You are now signed in

dependence on large amounts of demographic data when behavioral

human biases and errors in a data-rich environment

One of the most powerful uses is micro-segmentation based on behavioral characteristics

Digitization, more broadly, is also progressing unevenly among

Assets, including infrastructure, connected machines, data, and data

Operations, including processes, payments and business models,

The workforce, including worker use of digital tools, digitally-skilled

You are now

Feel free to select your topics of interest or change your

You are now signed in

There are also disparities between sectors in terms of degree of digitization:

In the United States, the information and communications

This unevenness can also be observed across countries; all have

The US economy as a whole is reaching only 18% of its digital

France has achieved 12% of its digital potential, the European

Emerging economies are even further behind, with countries

Digitization is transforming globalization, creating

You are now signed in

flows of information, searches, communication, video, transactions, and intracompany

The next wave of opportunity

You are now signed in

Advances in robotics, AI,toand

Machine-learning algorithms have progressed in recent years,

Computing capacity has become available to train larger and more

Massive amounts of data that can be used to train machine learning

You are now signed in

The combination of these breakthroughs has led to spectacular demonstrations like

Harnessing these evolving technologies will unlock multiple

Google has applied artificial intelligence from its DeepMind machine

You are now signed in

You are now signed in

AI and Automation will provide a much-needed boost to global

What about employment and work?

You are now signed in

You are now signed in

You are now signed in

Technical feasibility of automation, a critical first step that will

Cost of developing and deploying solutions;

Labor market dynamics, including supply and demand, and costs of

Business and economic benefits, not merely labor substitution

You are now signed in

A useful analogy to consider is that electric vehicles were demonstrated to be technically

You are now signed in

You are now signed in

You are now signed in

What should leaders do?