Professional Documents
Culture Documents
The Coming AI Revolution
The Coming AI Revolution
The Coming AI Revolution
In this report
The next frontier of innovation Shifting into high gear with intelligent automation
Retail and consumer products organizations are Brands and retailers are already adopting AI-powered intelligent automation at a
entering a new phase of technological innovation breathtaking pace – and that process is about to accelerate. Over 80 percent of
– with intelligent automation at its core. The executives in both the retail and consumer products industries expect their companies to
ramifications are profound, offering a host of be using intelligent automation by 2021, according to our new IBM Institute for Business
previously unimaginable capabilities – from Value study, developed in collaboration with the National Retail Federation.
automatically rerouting shipments to bypass bad
What’s more, 40 percent say their organizations are already engaged in some form of
weather, to personalizing in-store services based on
intelligent automation. Companies that aren’t experimenting with this capability risk
analysis of a customer’s facial expressions. Our
falling behind and need to move quickly if they hope to remain competitive.
newest research shows that two in five retailers and
brands are already using intelligent automation, and Why the surge in participation? Intelligent automation represents a major technological
that number is on track to double within the next breakthrough that has the potential to not just improve, but to transform the way
three years. Future AI-driven applications look to be companies do business. In intelligent automation, artificial intelligence (AI) is infused
far more extensive than many consumer products into automation, enabling machines to learn and generate recommendations, and to
and retail companies currently realize. make autonomous decisions and self-remediate over time (see sidebar on page 2,
“Intelligent automation and AI”).
85% of retail and 79% of To understand how brands and retailers are using intelligent automation today and what
consumer products companies they expect its future impact to be, we conducted a survey of 1,900 consumer products and
plan to be using intelligent retail executives who are leaders in the areas of supply chain and operations, and customer
automation for supply chain engagement in 23 countries. We then took a deeper look into impacts within industries and
planning by 2021 across organizational functions to determine how retailers and brands can address the
upcoming challenges and opportunities intelligent automation creates.
79% of retail and consumer
products companies expect to be We found that retail and brand executives have high expectations that intelligent automation
using intelligent automation for can boost their organizations’ bottom lines. Survey respondents anticipate that these
customer intelligence by 2021 capabilities can help reduce operating costs by an average of up to 7 percent, while increasing
annual revenue growth by up to 10 percent – four times the average revenue growth in
Retail and consumer products consumer products for 2017 and two times the forecasted growth in retail for 2018.1
executives project that intelligent
automation capabilities
Intelligent automation and AI
could help increase annual
revenue growth by up to 10% For our survey, we defined artificial intelligence as a capability in machines to reason,
remember information, learn and identify new insights through data discovery. Intelligent
automation is guided by AI tools that need minimal manual routine interventions. This
operational shift augments and assists human capabilities, reduces human errors and
builds efficiencies, while enabling digital operations and innovations. Four components
make up intelligent automation: the first three are fueled by AI, the fourth by automation.
Engagement over external touch points where users interact with systems
Learning from analytics across different data sources and recognizing semantic refer-
ences to use as criteria for decisions
Reasoning from learning to make autonomous decisions and self-remediate over time
Doing or executing to carry out the next best action that systems can execute digitally
and/or that people or robots can execute physically
3
Today, retail and consumer products organizations primarily use intelligent automation to “We’ve learned a lot about our
perform discrete internal processes that rely on existing rich-data sets, such as demand
forecasting and customer intelligence. But within the next three years, executives plan to
customer preferences from the
incorporate intelligent automation into more complex processes that require broader rich data we’ve collected over
sets of data, external collaboration and additional system integrations. And during that the past several years. Now, we
time, the projected penetration is expected to burgeon to more than 70 percent across
are in a position where we can
organizational areas that span the value chain (see Figure 1).
use data science and machine
Figure 1
More than 70 percent of retail and consumer products executives expect their companies to be engaging in learning to provide truly
intelligent automation across the value chain by 2021
personalized experiences
Retail Consumer products for our millions of customers.”
85% 86% Bindu Thota, Vice President of Technology
Supply chain planning Product design and development
zulily, LLC
85% 81%
Demand forecasting Manufacturing
79% 81%
Customer intelligence Demand forecasting
75% 79%
Marketing, advertising and Customer intelligence
campaign management
79%
73% Supply chain planning
Store operations
77%
73% Production planning
Pricing and promotion
4 The coming AI revolution in retail and consumer products
How retailers and brands use intelligent automation is changing – and the highest growth
is not necessarily in the areas one would expect. Over the next three years, we anticipate
seeing adoption surge most in areas that differ from those with the highest penetration
today. And projected adoption rates also vary between retail and consumer goods
companies based on each industry’s unique business requirements (see Figure 2).
Figure 2
The expected areas of highest growth in intelligent automation adoption over the next three years differ
by industry
49% 52%
41%
41% 57%
41% 43%
43%
40% 40% Piloting or
using today
Planning
37% within three
36% 33% 34%
29% years
27% 25% 26%
22% 21%
Supply Transporta- In-store Store Cross- Manufac- Product Product Supply Cross-
chain tion and services and operations functional turing design and lifecycle chain functional
planning logistics experience collaboration develop- manage- planning collaboration
ment ment
5
In designing and developing products, brands must consistently come up with new – and
hopefully trendsetting – design concepts (see sidebar, “AI search engine could inspire
next trend in fashion design”). To that end, brands can use intelligent automation
capabilities to ingest vast pools of data related to product use, as well as contextual and
global consumption information.
6 The coming AI revolution in retail and consumer products
They can then analyze and learn from that data to generate precise, relevant insights
Nike automates process for and apply them to designing products at unprecedented speeds. Some companies are
customer-designed sneakers5 already automating parts of their design and production processes, enabling customers
Nike Inc., an athletic footwear, apparel and to interact directly with systems to make and execute product design choices
equipment company, has developed a system (see sidebar, “Nike automates process for customer-designed sneakers”).
with which customers can design their own shoes
Targeting the supply chain and store operations in retail
and leave the store wearing them. With the new
Many of the retail executives in our survey, on the other hand, are exploring ways to apply
automated system, dubbed The Nike Maker
intelligent automation to cross-functional collaboration and interactions with customers.
Experience, customers put on blank Nike Presto X
These activities require more complex processes that involve additional system
shoes and, using voice activation, select graphics
integrations. This focus is evident in the two areas that show the highest growth in
and colors with which to design them.
intelligent automation adoption over the next three years: supply chain planning and
The system uses augmented reality, object in-store operations.
tracking and projection systems to display the
Supply chain planning involves collaboration across multiple functions, such as materials,
designed shoes for the customer. Once the
distribution and transportation planning. Previously, many of the processes tying these
customer has completed his or her choices,
planning functions together were manual.
the system prints the design on the sneakers,
which are available to the customer in under Intelligent automation is ideally suited for this type of environment. AI-powered tools can
two hours. The standard customization absorb data from different planning functions, and digest and analyze it quickly. They can
process can take up to two weeks. then produce calculations to help retailers make near real-time decisions when
developing and balancing plans, determining tradeoffs and gaining consensus. As they
work through the process, retailers can use automation to execute repetitive tasks, direct
workflows and execute resolutions to exceptions.
7
Store operations and in-store services can also benefit greatly from intelligent “Like all retailers we see the future
automation. Every city or neighborhood is unique, with its own highly localized flow of
people, places and events that shape consumer behavior and demand. A store in a
is in knowing who we are talking
college town requires different product assortment than a store in a resort area. to and speaking to them in the
Intelligent automation can learn from local data to determine products and services that right way.”
serve the needs of the neighborhood. Based on local venue characteristics and available
ingredients, it can automate assortment selection for a particular store. CMO, North American Retailer
AI technology also can apply what it learns to tailor in-store product and service offerings
to the individual customer’s needs. Imagine, for example, that you walk into a sporting
goods store looking for golfing gear. As you enter the store and opt-in for assistance,
the store’s AI-powered app accesses data about your purchasing patterns, interests
and preferences. It then automatically assigns you a sales associate who is a
competitive golfer.
At the same time, the app provides your information to the sales associate, so she is
equipped with pertinent knowledge at her fingertips. She greets you personally, strikes
up relevant conversation while leading you to the golf section of the store, provides
product-specific advice based on her golfing expertise and offers recommendations for
the right gear.
8 The coming AI revolution in retail and consumer products
The upsurge in intelligent automation and AI can offer unexpected benefits to brands and
retailers. Executives whose companies are using intelligent automation today are
experiencing a much higher degree of positive impact than the degree of impact the
executives in the planning stages expect (see Figure 3). And the order in which the two
groups of executives rank those impacts also indicates notable differences between
expectations and reality.
Figure 3
Executives implementing intelligent automation are seeing more transformational impacts
Improve quality and speed of decisions 54% 3 79% Increase operational efficiency
Reduce risk by improving visibility and processes 46% 7 66% Improve customer experience
Enable insights from integrated data sources 44% 8 65% Reduce risk by improving visibility and processes
Enable insights through enhanced analytics 43% 9 64% Enhanced insights from integrated data sources
The focus is less on lowering costs and more on increasing competitiveness and
sustaining long-term growth. While gaining efficiency and reducing costs may be the
initial impetus for engaging in intelligent automation, much greater benefits can be
unlocked as the capability matures.
98
Improve customer/consumer
79
business performance,
Improve enhance operational effectiveness or enable insights.
inventory productivity
86
107
By enhancing the customer experience, retailers can unleash entirely new approaches to
customer engagement and interaction. With intelligent automation, they can identify Retailer uses intelligent automation
customers’ anticipated needs at precise times and capture the right moment with the to speed up campaign development7
right offer in the pursuit of competitive advantage. A North American retailer had a wealth of
information on customer preferences and
For example, Avenue Stores LLC, a retail apparel chain, integrates data across multiple
buying habits. Because integrating this data
touchpoints, including in-store activities and market trend analysis, to learn and reason
into the campaign management tool was
about what customers want and when they want it. The company leverages real-time
difficult, the data went unused. The company
messaging and automated personalized offers to keep visitors engaged and incentivizes
wanted to harness this data and automate
them to purchase while they are in the right “shopping mode.”6
campaign workflows, making it easier to tailor
The automation of customer experience processes is seeing a little less traction content to the right audience. Marketing used
compared to other components of intelligent automation. Today, brands and retailers to require four days to design and launch
have begun to leverage AI-powered engines to automatically trigger email campaigns generic emails. Now, with AI-powered
(see sidebar, “Retailer uses intelligent automation to speed up campaign development”). campaign automation software, teams
An even more powerful use of this capability would be to apply it to the order fulfillment produce rich, personalized campaigns in just
process, enabling customers to make purchases directly from within the campaign. one and a half days, enabling them to respond
Improving inventory productivity and operational agility quickly to business needs. Marketing has also
Within the functions of supply chain planning, operations and logistics, Early Adopters boosted email open rates, helping incentivize
attribute the highest levels of impact to inventory productivity and operational agility. conversion and strengthen loyalty.
Both of these areas are critical to growth.
The goal of supply chain planning is to match supply to demand – delivering products to
the right place at the right time to meet customer needs and desires. Overstock can result
in markdowns and understock can mean lost sales. Either scenario has a negative impact
on inventory productivity and margin.
12 The coming AI revolution in retail and consumer products
To increase inventory productivity, retailers and brands can harness AI and machine
learning algorithms to optimize demand forecasting and supply chain planning. The
automation engine can then direct workflow to automatically update and adjust plans,
addressing demand spikes and dips due to both unexpected and planned events, such as
new product introductions or seasonal shifts.
For instance, The Procter & Gamble Company, a multi-national consumer goods
organization, is working toward deploying a next-generation demand planning solution to
improve its near-term forecast accuracy globally. Its goal is to enhance productivity and
equip planners to make better decisions in areas that are traditionally challenging, such
as promotional lift predictions.8
Operational agility – the ability within operations to adjust to change quickly and easily – is
another area in which intelligent automation can be a game changer. Companies are
increasingly using IoT and other smart devices to monitor their fleets and warehouse
operations, and to track the movement of goods throughout the supply chain. AI and
machine learning can apply learning and reasoning capabilities to monitoring and tracking
data. They can then use that data to derive insights and develop recommendations for next
best actions for handling potentially costly bottlenecks and damages.
In the case of bad weather or road conditions, systems equipped with intelligent
automation can dynamically reroute shipments to avoid delays and rebalance stock
levels as needed. For example, a new United Parcel Service machine learning app can
identify and eliminate bottlenecks based on cost-and-benefits analysis. If the app
detects an incoming storm, for example, it can reroute packages away from trouble spots
in a cost-effective way.9 By automating these processes using advanced algorithms,
companies can be better equipped to meet customer demand, maintain service levels
and handle unexpected events with agility.
13
While intelligent automation provides retailers and brands tremendous potential for Figure 5
Retail and consumer products executives indicate the right
transformational growth, executives in our study identified a number of critical factors to
skills and culture are critical success factors for intelligent
realizing its benefits. Key among them are obtaining the right skills, culture, infrastructure automation
and technology (see Figure 5).
Obtain the right skills and resources
Many companies have used AI to automate processes, but those that deploy it mainly to 43%
displace employees are expected to see only short-term productivity gains, according to
Create a culture open to change and adaptation
a Harvard Business Review report.10 The report found that the firms that achieve the most 41%
significant performance improvements are those that have successfully adopted
Align strategy with execution plans
intelligent automation to work alongside their employees.11 39%
Our study results indicate executives are in alignment with this way of thinking: While not Secure the necessary platforms and devices
38%
all roles are projected to change as a result of intelligent automation, 81 percent of survey
participants expect to reskill and retrain employees as they implement these capabilities Communicate a clear vision and benefits across the enterprise
30%
into specific functional areas. This could include training employees how to use customer
insights so they can offer more personalized services or equipping merchandisers with Ensure employees understand this capability
30%
tools that allow them to create more targeted inventory plans.
Ensure decision makers trust the automated decisions
In addition, executives expect intelligent automation to drive demand for new roles 26%
within their organizations. The top new roles cited by survey respondents include:
machine-learning engineers, data analysts, data scientists, AI interaction designers and
AI engineers.
14 The coming AI revolution in retail and consumer products
Figure 6 Online retailers are growing their footprints by opening physical stores, bringing fresh
Retail and consumer products executives name new risks
concepts and new energy to compete with traditional retailers. Both online and
associated with machine responsibility as a top challenge
traditional retailers aim to win over customers with compelling store experiences.
New categories of risk tied to machine responsibility In fact, over 70 percent of retail executives surveyed say they plan to implement
46%
intelligent automation in store operations and in-store services
Integrating capabilities with existing processes and systems
44% Consumer products and retail executives identify new categories of risk tied to machine
responsibility as the top challenge to adopting intelligent automation (see Figure 6).
Resistance in adapting to new capabilities
34% Working with intelligent automation relies on training employees and entering
Lack of skills and resources to execute effectively
high-quality data into the models over time. Without proper care, organizations can
34% unintentionally create biases that generate negative outcomes.
Alignment of vision to execution plans Consider the data that brands and retailers use for training, for example. Organizations
26%
using facial recognition technology to measure customer responses need to select
Lack of trust in automated decisions culturally diverse sets of images. While emotional expressions are largely universal,
26%
they do sometimes vary across cultures.
Lack of clear business benefits
15%
15
Recommendations
Work differently: Rethink the way work gets done in the digital age
–– Envision the end results. Be open to uncovering new capabilities and iteratively
evaluate automated tasks and activities for opportunities to redesign processes using
intelligent automation capabilities.
–– Partner with intelligent automation experts with design thinking expertise to guide and
host regular work sessions with executives and employees. Extend the invitation to
your customers, and allow your brand enthusiasts to co-create the next-generation
customer experience with you.
–– Think beyond basic automation to capitalize on the potential of intelligent automation.
This capability isn’t just about removing human activities from business processes, but
rather shifting to a culture of speed, agility and innovation.
17
Key questions
–– How can you better understand intelligent automation capabilities, and what steps is
your organization taking to communicate its capabilities and implications across the
value chain?
–– How do you plan to educate, train and prepare executives and employees to adopt this
capability?
–– What steps can you take to assess, identify and prioritize the strongest opportunities
for using intelligent automation to engage with consumers and improve the customer
experience, while ensuring back-end processes across the supply chain can deliver the
brand promise?
–– How do you plan to integrate intelligent automation across your existing tools to
engage with consumers and processes, and deliver the products and service
consumers demand?
–– How do you plan to build an enterprise approach for intelligent automation that can
deliver compelling brand experiences throughout the customer journey?
18 The coming AI revolution in retail and consumer products
Methodology
The IBM Institute for Business Value (IBV) and Oxford Economics surveyed 1,900 retail and
consumer products leaders in 23 countries between July and September of 2018.
North America
8% 5% 12%
5% UK and Ireland
5%
13% 7% France
7% Germany
Business 39% Region 7% Switzerland
functions 7% countries Italy
distributions distribution 7% Nordics
7%
Benelux
7%
39% 7% Japan
7% China
7% 7%
Australia and New Zealand
Supply chain, operations and logistics India
Mexico
Marketing and customer-focused areas
Chile
Merchandising, product design and development
Brazil
Finance and corporate functions
19
Related reports Lance Tyson is an Associate Partner for Supply Chain and Retail Operations within IBM
Schwartz, Robert, Kelly Mooney and Carolyn Heller Global Business Services. Lance is a senior level executive with 15 years of multi-channel
Baird. “The AI-enhanced customer experience: merchandising, planning, allocation and marketing experience best known for leading
A sea change for CX strategy, design and process change initiatives designed to tackle the challenges of today’s ever-changing and
development.” IBM Institute for Business Value. growing digital landscape. Lance can be reached at lltyson@us.ibm.com and on LinkedIn at
March 2018. https://www-935.ibm.com/services/ linkedin.com/in/lance-tyson-2966287/.
us/gbs/thoughtleadership/custexperience/ai-cx/
Christopher K. Wong is the Vice President of Strategy and Ecosystem for the IBM Global
Chao, Gene, Elli Hurst and Rebecca Shockley. Consumer Industry. He is responsible for setting the direction for IBM clients in the retail
“The evolution of process automation: Moving and consumer packaged goods industries. Chris has more than 20 years of experience in
beyond basic robotics to intelligent interactions.” areas ranging from sales and product management to corporate transformation. He led the
IBM Institute for Business Value. January 2018. IBM internal deployment of marketing technology, including the world’s largest B2B
https://www-935.ibm.com/services/us/gbs/ deployment of campaign automation and data systems. Christopher can be reached at
thoughtleadership/ibvprocessautomation/ cwong@us.ibm.com and on LinkedIn at linkedin.com/in/christopher-chris-wong-33048b1.
Butner, Karen, Dave Lubowe and Grace Ho. Notes and sources
“The human-machine interchange: How intelligent 1 IBM Institute for Business Value research; Ellet, Linda K. “Growth among consumer goods
companies continues to slow.” KPMG website. June 14, 2018. Accessed December 10, 2018.
automation is changing business operations.” IBM https://home.kpmg.com/uk/en/home/media/press-releases/2018/06/growth-among-consumer-
Institute for Business Value. October 2017. https:// goods-companies-continues-to-slow.html; Anders, Melissa. “Retail Industry Expects More Sales
www-935.ibm.com/services/us/gbs/ Growth In 2018.” Forbes. February 8, 2018. https://www.forbes.com/sites/
melissaanders/2018/02/08/retail-industry-poised-for-more-sales-growth-in-2018/#6eead5da4fa7
thoughtleadership/humanmachine/
2 Kushmaro, Philip. “5 ways industrial AI is revolutionizing manufacturing.” CIO website. September
27, 2018. https://www.cio.com/article/3309058/manufacturing-industry/5-ways-industrial-ai-is-
revolutionizing-manufacturing.html
3 Ibid.
21
4 Matchar, Emily. “Artificial Intelligence Could Help Generate the Next Big Fashion Trends.” © Copyright IBM Corporation 2019
Smithsonian.com. May 3, 2018. https://www.smithsonianmag.com/innovation/
IBM Corporation
artificial-intelligence-could-help-generate-next-big-fashion-trends-180968952 New Orchard Road
Armonk, NY 10504
5 Pandolph, Stephanie. “Nike’s new tech creates custom sneakers in under 2 hours.” Business
Insider. September 11, 2017. https://www.businessinsider.com/nikes-new-tech-creates-custom-
Produced in the United States of America
sneakers-in-under-2-hours-2017-9; Green, Dennis. “Nike has unveiled a new way to try on January 2019
sneakers at its stores without talking to anyone. Here’s how it works.” Business Insider. August 26,
2018. https://www.businessinsider.com/nike-scan-to-try-tech-how-it-works-2018-8 IBM, the IBM logo, ibm.com and Watson are trademarks of
International Business Machines Corp., registered in many jurisdictions
6 “Apparel chain Avenue digs deep to engage online shoppers.” STORES online magazine. December worldwide. Other product and service names might be trademarks of
5, 2017. https://stores.org/2017/12/05/the-new-way-to-know-your-customer/ IBM or other companies. A current list of IBM trademarks is available on
the web at “Copyright and trademark information” at: ibm.com/legal/
7 Schwartz, Robert, Kelly Mooney and Carolyn Heller Baird. “The AI-enhanced customer experience: copytrade.shtml.
A sea change for CX strategy, design and development.” IBM Institute for Business Value. March
This document is current as of the initial date of publication and may
2018. https://www-935.ibm.com/services/us/gbs/thoughtleadership/custexperience/ai-cx/
be changed by IBM at any time. Not all offerings are available in every
8 “P&G Adopts E2Open’s Demand Planning Tool Globally.” Consumer Goods Technology. February 2, country in which IBM operates.
2018. https://consumergoods.com/pg-adopts-e2opens-demand-planning-tool-globally THE INFORMATION IN THIS DOCUMENT IS PROVIDED “AS IS”
WITHOUT ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING
9 Woyke, Elizabeth. “How UPS uses AI to deliver holiday gifts in the worst storms.” MIT Technology
WITHOUT ANY WARRANTIES OF MERCHANTABILITY, FITNESS FOR
Review. November 21, 2018. https://www.technologyreview.com/s/612445/
A PARTICULAR PURPOSE AND ANY WARRANTY OR CONDITION OF
how-ups-uses-ai-to-outsmart-bad-weather/ NON-INFRINGEMENT. IBM products are warranted according to the
terms and conditions of the agreements under which they are provided.
10 Wilson, H. James and Paul R. Daugherty. “Collaborative Intelligence: Humans and AI Are Joining
Forces.” Harvard Business Review. July-August 2018. https://hbr.org/2018/07/ This report is intended for general guidance only. It is not intended
collaborative-intelligence-humans-and-ai-are-joining-forces to be a substitute for detailed research or the exercise of professional
judgment. IBM shall not be responsible for any loss whatsoever
11 Ibid. sustained by any organization or person who relies on this publication.
The data used in this report may be derived from third-party sources
and IBM does not independently verify, validate or audit such data. The
results from the use of such data are provided on an “as is” basis and
IBM makes no representations or warranties, express or implied.
87022787USEN-00