WHITEPAPER :

ARTIFICIAL
INTELLIGENCE
IN RETAIL
The Emerging Role of AI in Store Execution
 CONTENTS

01 The Intersection of AI and Retail

02 Customer Experience Depends on Retail Execution

03 Artificial Intelligence and Retail Execution

10 Looking Forward

11 Conclusion

12 Sources

00
THE INTERSECTION
OF AI AND RETAIL

In 1956, Dartmouth College launched the to grow from $2.9 billion in 2021 to $17.1
field of artificial intelligence (AI) as we billion in 2028.²
know it. After only six decades, AI under- Retail was a complex business before
pins most aspects of daily life, and retail is the pandemic and has only become more
no exception. so since. AI promises to solve, or at least

THE INTERSECTION OF AI AND RETAIL

We encounter AI dozens of times every dramatically improve, issues such as:
day, likely without realizing it. Anyone with • The increased pressure to develop
a smartphone uses AI when they ask Siri and maintain an omnichannel presence
about the weather or Google for the best • Persistent supply chain issues, starting
route to their destination. When you click with raw goods and manufacturing,
on a “recommended for you” link on a shop- and rippling through all layers of commerce
ping website, that link is generated by AI. • The need for more flexible, decentralized
AI has become so firmly established in work that fits the available retail
society that we can’t go back without labor marketplace
significant market disruption. The global All of these factors have culminated in a
market for AI is projected to grow from $84 perfect storm of miscommunication,
billion in 2021 to $1.6 trillion by 2030,¹ and stockouts, and employee turnover. AI,
AI drives much of the innovation and however, is a beacon of light guiding retail
optimization in the retail industry. In fact,
01
teams back to shore.
the global retail AI market is projected
 A 2018 Capgemini survey revealed that 74
percent of AI use cases serve customer-fac-
ing projects, while only 16 percent serve

CUSTOMER
operations.7 This heightened focus on
customer experience overlooks the founda-
tion of in-store success: retail execution.

EXPERIENCE
Everything a retailer does depends on a
customer finding the right product in the
right place at the right time. Consequently,
behind-the-scenes operational execution

Depends on
has just as much, if not more, impact on CX
as any consumer-facing effort.
Consider Walmart’s approach. In 2019,

Retail Execution Walmart equipped a neighborhood market

in Levittown, NY, with interactive displays
and a massive data center known as the
Intelligent Retail Lab. The IRL doesn’t
employ robots; instead, it uses AI-enabled
cameras to monitor the 50,000-square-foot
store, compare the quantities on-shelf to
Around the time the Dartmouth gang the projected sales demand, and trigger
cooked up the foundation for artificial intel- notifications to alert associates in real-time
ligence, the consumer marketing world what to restock and when, with granularity

CUSTOMER EXPERIENCE DEPENDS ON RETAIL EXECUTION

began to shift focus from the product to the down to the part number.8
customer, ultimately leading to the ascen- Walmart leverages AI to enhance
dance of customer experience (CX) as a customer experience, but they have a clear
leading indicator of in-store success. focus on automating in-store execution,
In a 2017 survey on AI implementation and for good reason.
priorities, 55 percent of the North American
retailer respondents indicated their top AI
priority was to optimize the customer expe-
rience.³ A 2020 MIT Technology Review
Insights survey focusing on the consumer
goods industry found a similar concern,
noting that the top AI use-case priority (48
percent) was customer care.4
In 2016, Lowe’s introduced the LoweBot, a
roving kiosk on wheels that engaged
customers via voice or touchscreen.5 In
2019, Giant Food Stores and Stop & Shop
deployed a service robot called Marty
across 200 stores to run price checks and
report aisle hazards.6 People, and the press,
often get excited about robots, but these

02
retail AI projects fail to address a crucial
piece of the retail value chain.
ARTIFICIAL INTELLIGENCE
and Retail Execution

While services like Buy Online Pickup

In-Store (BOPIS) have added new use cases
to the retail work day, the basic challenges
of retail execution are easily quantifiable
and measurable:
• How much of each product is on the
shelf/in storage in the back?
• When should product be restocked?
• Did a new display get built to
planogram specifications?
Retail teams on all sides of the aisle can
benefit from the added efficiency of AI, from

ARTIFICIAL INTELLIGENCE AND RETAIL EXECUTION

store managers tracking shelf status to
field reps verifying planogram compliance.
According to Juniper Research, retailers, While these various use cases are often
service providers, CPGs, and distributors lumped under the term artificial intelligence,
will spend $7.3 billion on AI by 2022, com- AI is an umbrella term that encompasses a
pared with the $2 billion spent in 2018.9 plethora of algorithms. Traditionally, AI
Some surveys indicate that AI in retail is refers to what is known as artificial general
quickly moving toward ubiquitous adoption intelligence (AGI), which describes a
at all levels of brick-and-mortar retail. system that replicates the functionality of
the human brain. At the 2009 Conference on
Some benefits of retail-execution-focused Artificial General Intelligence, an optimistic
AI include: 10 percent of the participants thought we
• Scheduling automation would achieve AGI by 2022. Half of the
• More accurate labor forecasting participants thought it would not be possi-
• Quicker response to volatile changes in ble until 2040, and 90 percent thought it
customer traffic and demand would take until 2075 to achieve AGI.
• More efficient and effective While AGI makes for interesting cocktail
inventory management conversation, practical AI promises profit-
• More empowered front-line workers ability without all the disturbing byproducts
• Streamlined compliance of the singularity. 03
 Systems that learn from experience. This area comprises
the bulk of real-world applications of AI currently in use in
MACHINE LEARNING (ML) retail. The algorithm is trained on real-world data to
replicate decisions performed by humans.

ML trained on large data sets over an extended period

DEEP LEARNING (DL)
of time.

ARTIFICIAL INTELLIGENCE AND RETAIL EXECUTION

Systems that use models of human neural networks
ARTIFICIAL NEURAL
to help computers learn. Commonly used in
NETWORKS (ANN)
conjunction with image recognition.

NATURAL LANGUAGE Systems that can process human language. Used most
PROCESSING (NLP) often for virtual assistants and chatbots.

Systems that identify and process human voice, such as

AUTOMATED SPEECH Siri, Alexa, Google, etc. Used for vocal commands and
RECOGNITION (ASR) speech-to-text applications.

A subsegment of AI (also known as Photo AI) in which

COMPUTER VISION systems interpret information from digital images, video,
and other media types.

04
AI is well-suited for in-store execution.
Retailers, manufacturers, and labor compa-
nies can use the collected data to work
smarter, observe shopper buying behaviors,
identify underutilized or over-tasked employ-
ees, optimize store-level execution, and
verify that shelves are neatly stocked with
purchasable merchandise.
More retail execution tasks suited for AI:
• Product availability
• Shelf inventory management
• Visual search
• Virtual assistants
• Payment services
• Supply chain management
• Fraud detection
• Predictive merchandising
• Programmatic advertising
• In-store visual monitoring
• Location-based marketing
The following use cases stand to provide
the largest and most immediate return on
investment for retail teams.

INVENTORY MANAGEMENT

ARTIFICIAL INTELLIGENCE AND RETAIL EXECUTION

Inventory management, sitting at the heart of
in-store execution, suffers from costly
challenges that have plagued retail teams for
decades. According to Gartner, only 18
percent of organizations report fulfillment
accuracy rates of 95 percent or higher.¹0
Considering that shoppers place on-shelf
availability (OSA) in their top three reasons for
choosing a store at any given time,¹¹ fractional
percentage improvements in inventory man-
agement (and corresponding reductions in
OSA) represent huge revenue increases.
Getting product to a store is only part of
the problem. Holding excess inventory (i.e.,
overstocking) results in a 32 percent increase
in costs annually.¹² Additionally, more than
50 percent of product categories fail to meet
planogram compliance at the shelf level,¹³
leading to an incalculable amount of lost
sales and compliance penalties. 05
 SMART SHELVES
Maintaining the ideal quantity of store Smart shelves, also known as intelligent
inventory, keeping shelves consistently shelving, utilize radio-frequency identifica-
stocked and in accordance with plano- tion (RFID) in conjunction with weight
grams—these costly inventory manage- sensors to track and report shelf inventory
ment challenges have haunted retailers and levels in real-time. As soon as a customer
suppliers since the dawn of modern retail, takes a product off a smart shelf, the shelf
and they are only growing more complex as sends that event data to the store’s data-
omnichannel shopping and supply chain base, helping store teams extract sales
complexities push the evolution of the insights, stay on top of inventory, and even
brick-and-mortar store. detect theft.
Store-level inventory management has Thanks to RFID, smart shelves have the
been a heavily manual, time-consuming, capability to interact with apps on shoppers’
and error-prone process. Store associates phones, opening the door to new customer
and service providers roam the aisles experience features. For example, shoppers
around-the-clock recording inventory can receive personalized promotions based
levels only to find their numbers outdated on past purchase history while they stroll
by the time they can take any meaningful the aisles, providing a digitized substitute to
action on them. Some of these teams traditional print promotions.
have employed mobile-first software to Because smart shelves rely on weight
replace antiquated, sluggish, paper-based detection, they are not a feasible solution
reporting, but many have yet to do so. for all product categories, especially appar-
Retail stores lose upwards of one trillion el and smaller-sized items, and cannot
dollars annually from out-of-stocks alone.¹4 assist in planogram compliance. Due to
Even mobile-equipped teams struggle to these limitations, retailers may not be able

ARTIFICIAL INTELLIGENCE AND RETAIL EXECUTION

keep pace with the volatility of store inven- to rely fully on smart shelf technology for
tory and the huge quantities of data that all inventory management needs.
mobile-execution tools generate. AI is
paving the way for more efficient, timely, ELECTRONIC SHELF LABELING (ESL)
accurate, and cost-effective inventory man- Electronic shelf labeling is a branch in the
agement processes. Smart shelf technolo- intelligent shelving family tree. ESL replaces
gy, electronic shelf labeling (ESL), and traditional print-based pricing labels with
photo AI (image recognition) are three digital labels that update instantaneously
particularly promising retail applications. and store-wide, eliminating the need for
costly and time-consuming labeling by hand.
ESL is a relatively straightforward technolo-
gy that provides substantial time-saving and
CX benefits to retailers.
Grocers, of all the retailer types, have the
most to gain from implementing ESL in their
stores due to the high product churn and
relabeling volume inherent with perishable
goods. According to Hussman, some
grocers spend upwards of $10 million
06
annually on price-changing labor, making reporting speed. With photo AI-equipped
them prime candidates for ESL.¹5 cameras in every aisle, retailers can track
While not technically a form of AI, ESL can both on-shelf stock and planogram compli-
integrate with AI systems to optimize in-store ance in real-time. However, the effective-
pricing. Product promotions, inflationary ness of shelf cameras comes at a cost, both
changes, and price matches become much literally and figuratively.
easier to manage when retailers combine Equipping every aisle of a store with
electronic labeling with artificial intelligence. advanced cameras can quickly become
With all these benefits, you may be won- expensive. When you multiply roll-out costs
dering why all stores haven’t transitioned to across a number of stores, the benefits may
ESL. The short answer: ROI. Rolling out ESL not outweigh the price for some retailers.
across a large number of stores is expensive Furthermore, aisle obstructions, either from
and laborious, and the ROI at a large scale is products or people, may inhibit shelf camer-
still uncertain, making many retailers as’ accuracy and result in skewed metrics.
hesitant to buy in. Aisle robots, on the other hand, may prove to
In the meantime, big-box retailers like be a more cost-effective application of photo
Walmart are experimenting with ESL in AI. While they cannot match the speed of
select locations to establish a proof of smart shelves and on-shelf cameras, they are
concept before scaling.¹6 Specialty retailers, still quicker, more accurate, and cheaper than
who can execute an ESL roll-out with much traditional labor-based inventory management.
greater ease and certainty, will likely lead the Many retailers have already begun rolling
charge in ESL adoption. out aisle robots of their own, including
Walmart, Giant Eagle, and Woodman’s to
PHOTO AI name a few. Sam’s Club got creative by
Photo AI, also known as image recognition, equipping their existing floor-scrubbing

ARTIFICIAL INTELLIGENCE AND RETAIL EXECUTION

utilizes computer vision and machine learn- machines with computer vision technology
ing to process and interpret in-store images from AI company Brain Corp. ¹7
of products. Like smart shelves, photo AI can Retailers clearly find value in aisle robots,
monitor and report on shelf stock without but the limitations of that value are not yet
the need for human review; but unlike smart understood. For now, and well into the future,
shelves, it can also analyze product place- retailers must use some combination of
ment and planogram compliance, further traditional labor for tracking inventory while
streamlining traditional in-store work. computer vision technology matures. During

360° RETAIL WFM SYSTEMS

Retail-focused image recognition hard- that maturation process, retail teams can still
ware falls into three categories: reap the benefits of computer vision by
1. On-shelf cameras supplementing photo AI technology into their
2. Aisle-scanning robots existing inventory management processes.
3. Mobile apps Many retail teams utilize mobile-based
On-shelf cameras are an exciting applica- retail execution systems to capture and
tion of photo AI because they are the closest share images of products in-store; these
substitute to intelligent shelving in terms of images (taken by employees) are sent to a
database for managerial review. Historically,
managers have manually reviewed hundreds

08
07
of thousands of these images a month, often
in sample sizes due to the overwhelming
volume, to verify shelf conditions. Photo-AI
integrations with mobile execution software
automate this painstaking and error-ridden
process, allowing teams to spend 33% less
time on inventory management, according to
an internal Movista study.
Whether through on-shelf cameras, aisle
robots, mobile apps, or some combination of
the three, photo AI provides immense bene-
fits to retail teams, and it will only become
more ubiquitous as computer vision matures.

LOGISTICS OF RETAIL EXECUTION

In-store teams aren’t the only retail constitu-
ents who benefit from AI. The logistics indus-
try is a prime candidate for AI, and many
enterprises have already jumped on board.
AI can track merchandise during transpor-
tation, reducing costs by identifying the most
efficient shipping process. Other AI applica-
tions help with real-time route optimization.
They track road and weather conditions and

ARTIFICIAL INTELLIGENCE AND RETAIL EXECUTION

give route recommendations, saving time
and money by reducing drive time.
Remember that Capgemini survey that said
74 percent of retail AI initiatives are customer
focused, while only 16 percent focus on oper-
ations? It might be time for retailers to
expand their use of AI beyond the front of the
store. The same study pointed out that using
AI to optimize the supply chain could save
retailers $340 billion annually.¹8 This is the
point where we move past the sexy, head-
line-grabbing aspects of AI in retail and drill
down into the execution details that have a
real, outsized impact on CX.

08
ROUTE AND SCHEDULE OPTIMIZATIONS OMNICHANNEL EXECUTION
Route optimization can improve the The retail store has become increasingly com-
efficiency and cost-effectiveness of store plex in recent years due to the merging of
visits. Effective route optimization helps physical and digital retail. Retailers and brands
you maximize completed orders and are trying to meet customers where they are,
deliveries by taking into consideration the and where they are is—well—complicated.
many factors that impact delivery perfor- Shoppers rarely buy products exclusively
mance, such as driver schedules, avail- through one channel (brick-and-mortar or
able hours, total stops, fulfillment e-commerce). In the vast majority of cases,
estimates, and legal requirements. Rout- shoppers fluctuate between physical and
ing isn’t about finding the shortest path digital depending on their needs and the type
from point A to point B; it’s about finding of product(s) they are purchasing. Brick-
the most efficient and profitable path. and-mortar retailers, eager to meet customer
The days of manually creating store expectations, have adopted BOPIS, curbside
routes and schedules are fading quickly. pickup, and direct-to-home e-commerce of
AI-powered route automation takes the their own. This evolution, though great for
guesswork out of route creation, helping customers, has created chaos among the
teams save time and maximize the efficien- various retail teams responsible for execut-
cy of store visits, no matter the scale. ing across these channels.
Store associates must find and retrieve
products off shelves for pick-up orders,
magnifying the intense need for better
shelf replenishment and monitoring prac-
tices. Brands must balance their invento-
ries across an even greater number of

ARTIFICIAL INTELLIGENCE AND RETAIL EXECUTION

channels, clarifying the need for better
business intelligence and real-time stock
management. And service providers must
navigate an increasingly complex store
ecosystem, exposing the need for better
cross-functional collaboration.
Thankfully, AI automation can help
resolve all of these pain points and more.
As the store continues to evolve, the
status quo manual processes and discon-
nected workflows become increasingly
inadequate. Retail teams have hyper-fo-
cused on using AI to enhance customer
experience, but the most valuable use
case for AI is keeping pace with rapidly
expanding customer expectations.

09
LOOKING FORWARD

There’s no stuffing the AI genie back into they had the data tools and systems to
its lamp. In our data-rich environment, AI implement AI across their organization to
enables retailers, suppliers, MSOs, and 55 percent.²0
other retail teams to operate at a level of A November 2021 AI software forecast
efficiency unachievable through tradition- from Gartner projects that worldwide AI
al manual processes. Recommendation software revenue will total $62.5 billion in
engines, a critical element of any digital 2022, an increase of 21.3 percent from
retail presence, run on AI. In physical 2021.²¹ The categories of knowledge man-
stores it takes on high-volume, tedious, agement, virtual assistants, autonomous
error-prone tasks such as pricing, invento- vehicles, digital workplace, and crowdsourced
ry management, and scheduling, freeing up data represent almost half of all spending.
front-line and back-office workers to focus According to our own projections, AI
on higher-value tasks. focused on retail execution provides
A recent Forbes survey showed that nine tangible improvements to inventory man-
out of ten leading businesses have invest- agement, in-store sales, and shelf compli-
ments in AI technologies, but only 14.6 ance, improvements such as:

360° RETAIL WFM SYSTEMS

percent deployed AI capabilities in their • 8 percent improvement in
work.¹9 The two main technologies in use on-shelf availability
are machine learning (e.g., planogram • 5 percent increase in sales

LOOKING FORWARD
compliance and price optimization) and • 10 percent improvement in compliance
autonomous machines (e.g., customer AI stands poised to move into new areas
service robots and product pickers). such as staff planning, recruitment (select-
Many retail teams are still trying to ing the right person for the right job), predict-
understand the best uses for AI in a phygi- ing product orders before they happen, and
tal (physical + digital) model. A Capgemini moving the orders up the supply chain faster
study of the 250 largest retailers by reve- than ever before. AI will become deeply
nue showed a year-over-year drop from
more than 80 percent who were confident
integrated within retail organizations and
affect every aspect of a business. 08
10
CONCLUSION
Artificial intelligence has come a long way since the 1956 Dartmouth conference. In everyday
life, AI is streamlining the ways we interact with the world. Getting directions is quicker and
easier than ever before; finding products that fit your style and needs is eerily seamless.
These enhancements to daily activities, however, may ironically be at the root of evolving
shopper expectations. AI’s saturation in society, in other words, is likely driving the need for AI
adoption in retail.
It is natural to assume that the most obvious AI use case for retail should revolve around
improving customer experience; after all, customers are the ones driving retail’s phygital
evolution. But deeper exploration reveals that the growing difficulty of retail execution far
outweighs the benefits of impressing shoppers with futuristic displays and no-check-out
stores (at least for the time being).
The best way to improve customer experience is by laying the foundation for a good
experience. Staying on top of inventory and shelf status, efficiently collaborating with internal
and external teams, and ensuring that shoppers get the product they want, where they want it,
when they want it—that is the foundation for a great shopper experience, and those are the
tasks which AI has the most potential to improve.
AI adds the most value to retail execution when it supplements or enhances existing labor
and operational strategies. AI that aims to replace the human element of the in-store custom-
er experience has an uncertain future, but AI focused on behind-the-scenes execution delivers
value to everyone involved.

Movista is a global, cloud-based retail execution and workforce management solution provider that
is transforming the future of work in retail. Movista stands as the world’s first and only platform to
enable collaboration between retailers, brands, service providers, and distributors, cutting costs and
lifting revenue for all retail stakeholders. Now, in-store and embedded teams can streamline work
and improve on-shelf availability. The Movista platform integrates with many critical business
systems such as Salesforce, SAP/S4HANA, Kronos, and Infor. CONCLUSION

For more information, visit www.movista.com.

SCHEDULE A DEMO
11
 SOURCES
Page 1
¹ https:www.globenewswire.com/news-re-
lease/2022/04/19/2424179/0/en/Artificial-Intelligence-Market
-Size-to-Surpass-Around-US-1-597-1-Bn-By-2030.html
² https://www.vantagemarketresearch.com/industry-report/ar-
tificial-intelligence-in-retail-market-1362

Page 2
³ https://www.statista.com/statistics/958052/retailer-priori-
ties-in-ai-implementation-north-america/
4 https://mittrinsights.s3.amazonaws.com/AIagenda2020/Glo-
balAIagenda.pdf
5 https://www.forbes.com/sites/blakemor-
gan/2020/05/13/the-
3-best-in-store-robots-and-why-they-work/?sh=4b0339437b22
6 https://www.washingtonpost.com/technology/2019/01/14/-
giant-food-stores-will-place-robotic-assistants-inside
-locations-company-says/
7 https://www.statista.com/chart/19351/ai-use-in-retail/
8 https://corporate.walmart.com/newsroom/2019/04/25/
walmarts-new-intelligent-retail-lab-shows-a-glimpse-in-
to-the-future-of-retail-irl

Page 3
9 https://www.forbes.com/sites/forbestechcoun-
cil/2019/03/29/why-retail-is-one-of-the-leading-sectors-investi
ng-in-ai/?sh=2d316fa1b4bd

Page 5
¹0 “Top Trends in Retail Digital Transformation and Innovation
for 2021.” Gartner. Accessed July 6, 2021
¹¹ https://www.fmi.org/forms/store/ProductFormPublic/solv-
ing-the-out-of-stock-problem
¹² https://www.fmi.org/forms/store/ProductFormPublic/solv-
ing-the-out-of-stock-problem
¹³ https://cupdf.com/document/planogram-compliance-mak-
ing-it-work.html?page=2

Page 6
¹4 https://www.ihlservices.com/news/analyst-cor-
ner/2018/06/worldwide-costs-of-retail-out-of-stocks/

Page 7
¹5 https://www.winsightgrocerybusiness.com/technolo-
gy/why-retailers-are-turning-electronic-shelf-labels
¹6 https://talkbusiness.net/2019/02/walmart-conducting-pi-
lot-test-of-electronic-shelf-labels-led-strips/
¹7 https://braincorp.com/newsroom/brain-corp-expands-com-
mercial-relationship-with-sams-club-to-power-in-club
-autonomous-robots-and-connected-data-services/

Page 8
¹8 https://www.statista.com/chart/19351/ai-use-in-retail/ SOURCES

Page 10
¹9 https://www.forbes.com/sites/gil-
press/2020/01/13/ai-stats-news-only-146-of-firms-have-deplo
yed-ai-capabilities-in-production/?sh=762b318e2650
²0 https://www.statista.com/chart/19351/ai-use-in-retail/

12
²¹ https://www.gartner.com/en/newsroom/press-releas-
es/2021-11-22-gartner-forecasts-worldwide-artificial-intelligen
ce-software-market-to-reach-62-billion-in-2022