7 leading machine

learning use cases

How today’s businesses are using machine learning
to achieve fast, efficient, and measurable results
I N T R O D U C T IO N

Solve your most common

business challenges with
machine learning

Machine learning (ML) has moved beyond the hype to become a meaningful Second, you’ll want to find a use case that is rich in data that you already have.
driver of value for many organizations. Over two-thirds of businesses that A good business use case with no data will lead to frustrated data scientists.
have fully embraced artificial intelligence (AI) say the technology has created
a better customer experience, and more than half say it has improved Lastly, you’ll want to evaluate whether your business problem actually
decision-making, increased productivity, and allowed for innovation while requires ML for success and whether ML will result in better outcomes
achieving cost savings.1 than your traditional approach. These outcomes might be realized as cost
reduction, increased employee productivity, or an improved experience for
While it’s clear that ML is an essential part of business transformation, many your customers.
organizations struggle to understand where to apply ML for the most impact.
Selecting the right ML use case requires you to consider a number of factors. The best way to satisfy all these criteria is to ensure that technical experts
and domain experts are working hand in hand on your ML project. Technical
First, you need to find a balance between optimal business value and speed. experts can conduct feasibility assessments, and domain experts will ensure
A proof of concept built by a siloed data scientist is not likely to generate the solution is solving a real business problem and will have a real impact.
much enthusiasm for ML in an organization. What is more apt to attract the
needed commitment and funding is showing how ML can address the practical
issues your organization currently faces. Furthermore, you’ll want to find
something that can be accomplished in 6–8 months so that you won’t lose
momentum. This is especially true if this is your first foray into ML.

“Artificial Intelligence (AI) and Analytics,” PwC Tech Effect Blog, 2022
2
1
Starting with the right use case
is key to organizational buy-in
In this eBook, we have outlined seven use cases in which Amazon Web Services (AWS) 7 leading use cases:
customers have successfully applied ML. These use cases can strengthen your business case
for wider adoption of ML, and you can apply them to kick-start your ML journey or add them
to your current strategy.
1 Improve employee productivity ›
What makes a good ML use case?
2 Automate document data extraction
and analysis ›

• Solves a real problem for your business—one that’s important enough to get attention, 3 Add AI to any contact center ›
support, and adoption
• Increases performance, reduces costs, or improves your customer experience
4 Improve customer self-service
experience ›
• Includes technical experts to conduct feasibility assessments and domain experts to
ensure the solution will be used 5 Personalize customer
recommendations ›
• Can be completed in 6–8 months
6 Automate content moderation ›
When you are ready to deploy your use case, you have the choice of using one or
more fully managed AWS AI services to quickly get started and integrate intelligence 7 Validate user identity ›
into your applications. Or, if you want to develop your own models, you can use
Amazon SageMaker—a solution that provides you with all the ML tools you’ll need
in a single service.

3
1. Improve employee productivity ID E A L F O R A LL I N D U ST R IES

by quickly and easily finding GILEAD SCIENCES, INC.

accurate information “Amazon Kendra is a turnkey AI solution that,
when configured correctly, is capable of spanning
Employees who have fast, easy access to accurate data are more productive. every single domain in the organization while
In a 2019 study by the Economist, executives identified “ease of access to being straightforward to implement.”
information required to get work done” as the number one way in which
Jeremy Zhang, Head of Advanced Analytics
technology can drive employee engagement and productivity.2

Your employees can search for the information they need by asking natural
language questions through Amazon Kendra—a highly accurate intelligent
search service powered by ML. This is much faster and more efficient than BAKER TILLY DIGITAL
traditional keyword search, and the service is easy to deploy for businesses “Amazon Kendra allows product managers to ask
of all sizes. The resulting boost in productivity helps accelerate research
questions in everyday language such as ‘What
while improving decision-making—and can strengthen your business case
for wider ML adoption. parts are made of titanium?’, quickly surfacing
an answer previously not possible with keyword
Learn more › search and connecting them to relevant content
across an enterprise-wide repository, or providing
marketing managers quick access to crucial
research on customer behavior.”

Ollie East, Director of Advanced Analytics & Data Engineering

Tom Puch, Sr. Manager

“The Experience of Work,” The Economist, 2019

4
2
2. Make faster decisions by ID E A L F O R
Financial Services, Healthcare and Life Sciences, Accounting,
automatically extracting and Education, Government, Legal, Oil and Gas

analyzing data from documents

ASSENT COMPLIANCE
The millions of documents created by your organization contain a treasure “Amazon Textract’s OCR technology enabled us to
trove of insights waiting to be leveraged. Unfortunately, manually processing
process…documents while Amazon Comprehend
the ever-growing volumes of data to make them easy to access and search
is a cumbersome, costly task. Using ML, your organization can gain timely
was able to extract custom entities. [Using] Amazon
access to the information contained in your documents, leading to new Augmented AI, we were able to have our teams
insights that inform your business decisions. review documents in a given accuracy range and
help train our next model iteration. Combining
For organizations looking to activate intelligent document processing today,
AWS offers three services that can be deployed individually or combined as these services…[saved] our customers hundreds
building blocks to develop an end-to-end document processing solution. of hours in manually reviewing documents.”
Amazon Textract automatically extracts handwriting, printed text, and
Corey Peters, AI/ML Team Lead
data from scanned documents. Amazon Comprehend is a natural language
processing (NLP) service that uses ML to find insights and relationships in
text. And Amazon Augmented AI (Amazon A2I) provides built-in human
review workflows to help ensure the accuracy of the data. THOMSON REUTERS
“Our solution required several iterations of
You can also quickly and efficiently develop your own ML models for
text extraction and analysis with SageMaker, a fully managed service deep learning configurations at scale. Amazon
that helps business analysts, data scientists, data engineers, and MLOps SageMaker enabled us to design a natural
engineers prepare data and build, train, and deploy ML models quickly. language processing capability in the context of
This service provides several built-in ML algorithms—such as BlazingText
a question answering application…successfully
and Linear Learner—that are optimized for text classification, NLP, and
allowing [our customers] to simplify and derive
optical character recognition (OCR).
more value from their work.”
Learn more › Dr. Khalid Al-Kofahi, VP of Research & Development

5
3. Add AI to any contact center to ID E A L F O R A LL I N D U ST R IES

improve service and reduce costs VANGUARD

“At Vanguard, we aim to build state-of-the-art
Improving the customer service experience is one of the best ways to
experiences for our clients. AWS CCI’s self-service
differentiate your brand—and demonstrate the value of ML. Successful
solution enables us to rapidly iterate and think
organizations treat their customer contact center as a business-critical asset
rather than viewing it solely as a cost center. about the client experience from a different lens.
This allows us to continue improving call routing
ML can help transform a contact center into a profit center by reducing call
accuracy, improve the odds of first call resolution,
wait times, improving agent productivity and satisfaction, lowering costs,
and helping to identify business improvement opportunities.
and deliver value for our clients and business.”

Nicki Larro, Retail Contact Center Technology Product Owner

AWS offers several flexible options to add intelligence to your contact
center. If you are looking for an end-to-end solution, Amazon Connect is an
easy-to-use omnichannel cloud contact center that helps companies deliver
superior customer service at a lower cost. MAGELLAN HEALTH
“We chose Amazon Kendra, a service within AWS
If you already have a contact center in place, AWS Contact Center
Intelligence (CCI) solutions empower you to improve customer experience, CCI to build a secure and scalable agent assist
boost agent productivity, and gain conversation insights by adding AI application. This helped call center agents, and
capabilities to the contact center of your choice—without any ML expertise. the customer they serve to quickly uncover the
AWS CCI solutions integrate with most leading contact center platforms like information they need. Since implementing
Genesys, Cisco, Avaya, Amazon Connect, Mitel, Talkdesk, and many more.
CCI…[solutions], early results show an average
AWS CCI solutions use a combination of AWS AI services to provide reduction in call times of about 9–15 seconds,
self-service virtual agents, real-time call analytics and agent assist, and which saves more than 4.4k hours on over 2.2
post-call analytics.
million calls per calendar year.”

Learn more › Brian Lichtle, Sr. Director of Software Engineering

6
4. Improve customer self-service ID E A L F O R A LL I N D U ST R IES

experience with conversational AI RYANAIR

“Ryanair’s chatbot powered by AWS conversational
Conversational AI (CAI) interfaces add human-like conversation capabilities
AI services has become an extension of the
to your business applications by combining different natural language
customer care team, handling over 3 million
technologies like NLP, natural language understanding (NLU), and natural
language generation (NLG). The demand for self-service conversational conversations across five languages every year.”
interfaces continues to grow as more and more users prefer to interact with
Lee Reddin, Head of Service Design
businesses on digital channels. Organizations of all sizes are developing
voice and text conversational interfaces to increase user satisfaction, reduce
operational costs, and streamline business processes.
XPERTAL
CAI interfaces are used broadly across a wide variety of industry segments
“Xpertal improved contact center efficiency,
and use cases. The common use cases for CAI include 1) Building virtual
omnichannel support across email, phone,
agents and voice assistants, 2) Automating informational responses and data
capture, 3) Boosting agent productivity in contact centers, 4) Automating collaboration tools, and internal corporate
customer service, and 5) Performing transactional operations. websites. With Amazon Lex’s easy to use interface,
our Contact Center team was able to create bots
CAI solutions primarily leverage Amazon Lex, which is complemented
by additional AI and ML services, including Amazon Kendra, Amazon
after a 1-hour training session. The solution was
Comprehend, Amazon Translate, Amazon Polly, and SageMaker. Amazon easy to integrate with business applications to
Transcribe combined with Amazon Lex provides the advanced deep learning create a virtual help desk that enables employees
functionalities of automatic speech recognition (ASR) and NLU to enable to find answers faster. Top benefits include
customers to build applications with highly engaging user experiences and
improved call deflection rates, reduced call wait
lifelike conversational interactions using voice and text. With Amazon Lex, the
same deep learning technologies that power Amazon Alexa are now available
times, and increased agent productivity.”
to any customer, enabling them to build sophisticated natural language Chester Perez, Contact Center Manager
conversational bots quickly and easily.

Learn more ›

7
5. Deliver personalized ID E A L F O R
Retail, Media and Entertainment, Travel and Hospitality, Education,
recommendations to increase Financial Services, Government, Healthcare, Software and Internet

customer engagement
WARNER BROS. DISCOVERY
Consumers today expect real-time, personalized experiences across digital “With Amazon Personalize we were able to build
channels as they consider, purchase, and use products and services. and train a real-time recommendation engine
ML can help you deliver these highly personalized experiences, resulting in
POC within two days. Since deployment on our
improvements in customer engagement, conversion, revenue, and margin. TBS, TNT, TruTV and Adult Swim web properties,
AWS offers ML solutions that deliver higher-quality personalized experiences over 25k unique consumers have clicked on cross-
across digital channels—all tailored to your business needs. portfolio promotions for the movies, shows and
To get started quickly with personalization today, you can use site sections recommended by Amazon Personalize.
Amazon Personalize—a fully managed service that leverages more than For the users receiving personalized promotions
20 years of personalization experience at Amazon. Amazon Personalize makes we have seen total user engagement increase by
it simple to develop applications for a wide array of personalization use cases, 14% and cross brand engagement increase by 12%
including product or content recommendations, individualized search results,
compared to a randomized control group.”
and customized marketing communications—with no ML expertise required.
Don Browning, VP Cloud Architecture
Or, if you want to develop your own ML models for recommendation
engines, you can use SageMaker—a fully managed service that helps ML
practitioners, including business analysts, data scientists, data engineers,
GEVME
and MLOps engineers, prepare data and build, train, and deploy ML models
quickly. You can use built-in algorithms, such as factorization machines or “AWS had all the tools we needed to create
XGBoost, both of which are optimized for personalization, to readily train next-generation virtual event software and
and deploy models. You can also bring your own algorithms and models or
networking opportunities.”
select from the hundreds of algorithms and pretrained models available in
Amazon SageMaker JumpStart. Veemal Gungadin, Founder & CEO

Learn more ›

8
6. Automate content moderation ID E A L F O R
Social Media, Gaming, Ecommerce, Advertising, Financial Services,
with AI to protect users, brands, Healthcare, and Education

and information
DREAM11
Eighty percent of all web content today is user-generated content (UGC). “AWS promotes a user-first culture, with intuitive
Seventy percent of consumers agree that brands are responsible for cloud-native services that help us launch things
moderating UGC, and 40 percent disengage with brands after a single toxic fast without any dependencies, develop and make
content exposure.3 As the volume, complexity, and speed of UGC increase,
our prototypes live very quickly, even at a massive
organizations of all sizes must commit significant technical and human
resources to ensure customers are not harmed by offensive material and that scale. This gives us a competitive edge in the
sensitive information is not exposed publicly, all while keeping operational market, where speed is essential.”
costs down. Human-based content moderation alone cannot scale, leading to
Praveen Jain, VP of Engineering
poor user experience, high moderation costs, and brand risk.

AI and ML can help organizations automate content moderation to reduce

human workloads. AWS has fully managed yet customizable AI solutions and COSTAR GROUP
services that can help you automate how you protect your customers, online
“Amazon Rekognition’s Content Moderation
communities, and brands while lowering moderation costs.
API enabled us to easily build a solution to
Amazon Rekognition helps you streamline or automate image and video automatically analyze all uploaded images,
moderation workflows to detect inappropriate or offensive content
allowing us to efficiently deliver high-value
and unwanted brand associations. Amazon Transcribe enables you to
convert speech to text, check it against banned words, and detect harmful
products and improve our product offerings by
audio-focused content. Amazon Translate helps moderate text content across making the images we receive more discoverable
languages, from social media comments, reviews, and documents to in-game and our community more inclusive.”
chats and instant messages. Amazon Comprehend increases your investment’s
value with contextual insights and protects sensitive personal and health Mark Osborn, Principal Software Engineer

information, and Amazon Augmented AI brings human moderators in the loop

to verify decisions and improve the ML model over time.

Learn more ›

“Survey: Nearly Half of Americans Quickly Lose Trust in a Brand If Exposed to Toxic or
9
3

Fake User-Generated Content on Its Channels,” Business Wire, 2021

7. Validate user identity to ID E A L F O R
Finance, Education, Ecommerce, Retail, Gig Economy
protect users and prevent fraud
INTER
The fraud detection and prevention market is expected to grow to $75 “Three years ago, we opened 200 accounts a day.
billion USD by 2028 at a 16 percent CAGR4 in an increasingly digital world
Today there are 29,000 accounts opened daily, and
where organizations have to invest billions of dollars to reduce and prevent
fraud. While identity verification can be used to prevent fraud, correctly
we would not have the agility to do this without
identifying user identity in real time is technically complex and Amazon Rekognition.”
resource-consuming and can add friction to the customer experience.
Bruno Picchioni, Squad Lead, Credit Platform
Verifying identity to combat fraudulent activities makes for a good AI and ML
use case for three primary reasons. First, ready-to-use AI solutions, services,
and ML models enable developers to quickly deploy identity verification AELLA CREDIT
processes. Second, fully managed APIs can help train and implement custom
“Using Amazon Rekognition for identity
models with less effort and data. Third, customers can leverage existing
development resources with solutions and services that require no previous verification on our mobile application has reduced
expertise, making AI and ML readily available across the organization at a verification errors significantly and given us the
lower cost. ability to scale. We can now detect and verify
an individual’s identity in real time without any
Amazon Rekognition Identity Verification offers pretrained facial
recognition and analysis capabilities that organizations can quickly human intervention, thereby allowing faster
integrate to authenticate their users’ identities. Through a streamlined access to our products. We tried various well-
verification process and the ability to detect fraudulent and duplicate advertised solutions, but none of the popular
accounts in seconds, onboarding legitimate customers becomes simple
alternatives could accurately map out various skin
without affecting your customer experience. Leveraging automation
tones. Amazon Rekognition helped us effectively
and AI with ready-to-use services, your organization can effectively
reduce implementation and operational costs. recognize faces of our customers in our markets. It
also helped us with KYC in discovering overlapping
Learn more › profiles and duplicate datasets.”

Wale Akanbi, CTO & Co-Founder

“Fraud Detection and Prevention Market Size Worth $75,139.66 Million, Globally,
10
4

by 2028 at 16% CAGR - Exclusive Report by The,” Bloomberg, 2022

Start or expand your
machine learning journey now
With the use cases in this eBook, you can leverage ML to boost productivity, make
smarter use of your data, meet customer demands more effectively and efficiently,
enhance customer experience and satisfaction, make better decisions faster, and
reduce the frequency and impact of fraud.

We chose to highlight these seven use cases because our customers are achieving
impactful, measurable results with them today—and because they fulfill the
requirements you should look for when identifying a suitable application for ML.
These use cases can be completed in a matter of months, solve real business problems,
increase performance, reduce costs, and improve the customer experience. They lend
themselves to the inclusion of technical and domain experts and—when properly
executed—generate results that gain attention and foster executive support for wider
ML adoption.

The business potential of ML goes far beyond these seven use cases. With the broadest
and deepest set of AI and ML services and tools available today, AWS can help you
apply ML in a wide variety of ways that can transform your business—allowing you
to push innovation to new heights and reimagine the possibilities of what your
organization can achieve.

Learn more about AWS AI use cases ›

©️ 2023, Amazon Web Services, Inc. or its affiliates. All rights reserved.