Automation Using Cloud-native Services

Whitepaper by Vinod Kumar Mohan, Automation Consultant, Senior Manager |

Veena S, Enterprise Automation Leader, Vice President | Renjeev Kolanchery, Enterprise Automation Head,
Senior Vice President
Contents

Introduction 1

Cloud-native Service-based Automation 1

How to Automate Business Processes with Cloud-native Services? 2

AI & ML – Artificial Intelligence and Machine Learning in Automation 4

IDP – Intelligent Document Processing in Automation 4

HIL – Human in Loop in Automation 5

Governance and Reporting in Automation 6

Why is Process Automation with Cloud Native Services Futuristic? 6

Use Case 8

What is the Difference Between Typical RPA Tools vs. Automation Using Cloud? 10

Conclusion 10
1.
Introduction
Today, most business wants to be tech-intensive to stay ahead of the curve and be future-ready.
In this expedition, enterprises are accelerating cloud transformation to drive value for the company
and bring in scalability, affordability and optimized pricing models for their day-to-day operations.
However, conventionally Process Automation was not built into the cloud; it requires expertise
and specialized tools. While few cloud-enabled automation platforms are readily available in the
market, they are expensive. Hence, it would be cost-effective and scalable to build the automation
framework using cloud-native services to automate the business processes for the future.

Mphasis has deep expertise in Process Automation and Transformation, leveraging multiple
technologies across RPA, intelligent data extraction, AI/ML algorithms and cloud-native platforms.
This experience helps to enable business process automation to bring down the automation cost
and help ease development and maintenance for the automation developers.

2.
Cloud-native Service-based Automation
Some advantages of moving to cloud-native services are stability, reliability and highly available
applications. However, despite these advantages, organizations face challenges while adopting
cloud-native technologies.

Here is the list of challenges that corporates face with the current set of automation platforms and
how these challenges can be mitigated with the cloud-native automation approach.

Cloud-native services are evolving,

No Proprietary source code and Choice of the Hyperscalers
cloud-native platform. Hence, customers
Limitations and there is a continuous contribution
Lock-in of the Tool from the tech community to handle
have complete flexibility over the choice of complex scenarios.
tech stack.

Skills like Selenium are easy to learn and Cloud hyperscalers manage
skilled people are also abundant in the market. the services, and the
Skill Maintenance customers have to do very
Selenium can be implemented in various
programming languages like Java, Python, little maintenance.
C# and JavaScript.

No licensing cost.
Mostly, pay-per-use
License High Cloud deployments are
of the cloud services. Cost Availability highly available, and
disaster recovery is baked in.
& DR

Very few cloud-based automation and cloud-native

Infrastructure cost is reduced as tools are available in the market. With the
you pay-per-use for most of the
Infra Cloud-based tremendous push by all companies to migrate to the
services on the cloud. Cost Automation cloud, cloud-based automation is one solution that
fits into the enterprise's technology vision.

| 1
3.
How to Automate Business Processes with
Cloud-native Services?
Mphasis works with multiple automation platform providers, has partnerships, an expert team
and a proven record of delivering significant value to customers.

Any typical automation platform needs to have these components to develop, test and deploy
the bots:
 Studio to build automation code
 Packaging and deployment architecture
 Execution framework for the bots
 Orchestration framework
 A control tower to manage the tasks
 Password management
 Datastores
 OCR and Human in Loop capabilities
 Machine Learning capabilities
 Logging/schedulers

Control Tower

SDK/API for Bot Execution OCR + Human in Loop

Framework for the Orchestration
Studio Development Framework + Machine Learning
Bots

Packaging & Development Password Manager File Storage/Datastores Logging/Schedulers

Security, Disaster Recovery, Upgrades

Product Support, Developer Community

Figure 1: Typical offerings of an automation platform provider

Cloud hyperscalers offer reliable, scalable and inexpensive cloud computing services. In addition,
they provide various services that are not limited to Compute, Networking, Storage, Database,
Analytics, Messaging, etc. This gives us the flexibility to put together services that can help build
a full-fledged automation platform.

The table below illustrates how to map the capabilities of an automation tool to cloud services.
Here, we have shown examples of Azure and AWS. This could be achieved with other cloud
hyperscalers too.

2 |
 Components Azure AWS

PyCharm, Microsoft Visual Studio, IntelliJ IDEA, NetBeans, Eclipse, AWS Cloud9
Studio
(Cloud-based IDE), Microsoft Visual Studio Online (Cloud-based IDE)

Selenium code using Java/Python/C#, Selenium code using Java/Python/C#,

SDK/API
JavaScript JavaScript

Execution
Azure Functions, Container Instances, Lambda, Elastic Cluster Service, Elastic
Framework for
Azure Kubernetes Service, VM’s Kubernetes Service, VM’s
the Bots

Orchestration Logic Apps, Azure Event Grid Event Bridge, Step Functions

OCR Azure Form Recognizer Textract, Rekognition

Content Moderator (Microsoft Cognitive

HIL Amazon Augmented AI
Services), Custom UI

Machine
Learning/Artificial Microsoft Cognitive Services Machine Learning Services (SageMaker)
Intelligence

Azure DevOps or any other combination AWS Code Build or any other combination
of ci-cd services. Cloud infrastructure can of ci-cd services. Cloud Infrastructure
Packaging &
be provisioned using Azure Resource can be provisioned using code templates
Deployment
Manager code templates and plugged using AWS CloudFormation (& SAM) and
into the DevOps pipelines. plugged into the DevOps pipelines.

Password Manager,
File Storage,
Standard services across cloud hyperscalers
Datastores,
Logging, Scheduler

Security,
All managed resources from the cloud hyperscalers ensure that these horizontals are
Disaster Recovery,
handled well
Upgrades

Product Support/ Cloud hyperscalers maintain the managed resources

Developer Developer Community – Selenium has been a tool of choice for test automation for a long
Community time and has a large developer base

Cloud-native landscape is vast, and it is easy to get lost in its expanding number of competing and
overlapping platforms and technology options. We help organizations transform their environment
by adopting cloud computing services and leveraging Artificial Intelligence (AI) & Machine Learning
(ML) powered apps in automation.

Automation scope is broadened, and efficiency increased with the inclusion of AI & ML. It is an
added layer of automation that brings human cognition into workflows. Our expertise and
market-ready products in AI & ML help us make automation more efficient.

| 3
4.
AI & ML – Artificial Intelligence and Machine Learning
in Automation
AI & ML can enhance the efficiency of automation and expand its possibilities. Cloud, with its
on-demand, highly scalable and high-performance computing capabilities, serves well for building
AI/ML models requiring intensive computing.

Cloud hyperscalers offer managed services that are purpose-built for use cases like Natural
Language Processing, Text Extraction from images, Document Classification, etc. They are built
using high-end AI/ML models and data that provide highly accurate results.

Some of the applications of AI & ML in automation are:

Process Mining
With the ever-growing number of business processes in an organization,
with multiple overlapping processes also sharing data, simplifying business
NLP-based Chatbots processes and identifying candidates to automate can be challenging.
Natural Language Processing-based AI/ML can help with process mining, analyzing logs and input/output business
Chatbots provide a natural language process data to mine for processes that can be analyzed by SMEs and filtered
interface that can be used to trigger for automation.
automation on the backend. This way,
a non-technical user can also interact
with the bots.

Document Recognition and Data Extraction

Any document processing automation would require
ITOps identifying the document type and extracting data from it.
AI/ML helps to build rules for classification and data
ITOps is a growing area where data from IT extraction by seeing only a subset of documents.
operations is analyzed for patterns and anomalies
detected to take pre-emptive, preventive actions.

All the above applications only enhance the overall benefits of automation, making the application
of AI/ML to automation a necessity and not an add-on.

5.
IDP – Intelligent Document Processing in Automation
Intelligent Document Processing is a specialized tool driven by Artificial Intelligence/Machine Learning.
It helps automate data extraction from complex semi-structured/unstructured documents and convert
them into structured, usable data. It is also referred to as Cognitive Data Processing or Intelligent Data Capture.

4 |
IDP takes advantage of Artificial Intelligence (AI), Machine Learning (ML), Optical Character
Recognition (OCR), Computer Vision and Intelligent Character Recognition (ICR) technologies to
classify, categorize, extract relevant data and validate the extracted data for improved accuracy.
Some of the applications of IDP for automation are:

KYC
Know your customer documents data extraction. This is a complex use case
in automation, as each country has several templates. In addition, these
documents are typically scanned, meaning quality is an issue – making data
extraction more difficult.

Accounts Payable
Manual processing of invoices is a tedious, error-prone task.
IDP can identify, extract and organize relevant data – freeing
up the manual human workforce.

IDP helps with hard-to-automate problems and can be valuable to the automation solution kit.

6.
HIL - Human in Loop in Automation
Despite the advances of AI/ML, there are always scenarios that these technologies cannot
handle to the extent desired. In such cases, having a human review can help close the gap.
Human reviews can also be included as random QC, increasing the confidence in such solutions.

Human in Loop refers to a data review by human workers, which is part of a workflow.
HIL is especially useful for validating and updating data processed by OCR and AI-ML tasks.

Human reviews can be triggered based on:

• Low confidence data extraction scores
• Missing fields information
• Randomly for QC

Business users/SMEs are provided with a UI Template (simple web application) where the data
is presented for validation and update if necessary. Authenticated access can be provided to the
users using any Open ID connect identity providers.

These reviews can be orchestrated as part of workflows, helping improve automation potential,
data quality and QC.

Managed services from the cloud hyperscalers provide Intelligent Data Extraction, AI-ML and HIL
platforms which can be brought together within an automation workflow.
| 5
7.
Governance and Reporting in Automation
Governance is a critical component to achieving standard outputs with scale in automation.
Reporting provides valuable insights into the performance of automation, which is vital in value
realization and justification of efforts. Together, they are important pieces of the automation process.

Some of the critical aspects are given below:

Access Monitoring and Observability

There are various users of an automation solution like Monitoring automation solutions deployed on the
developers, admins, business users, analytics teams, etc., cloud, understanding resource utilization and
and each user would require specific access and privileges. getting a unified view of the operational health
will help keep the solution running in good
Access to resources and services on the cloud is health. There are various tools in the cloud that
controlled using Identity and Access Management allow for monitoring, capturing metrics and
services provided by all cloud hyperscalers. taking action, if necessary.
In addition, Role-Based Access Control (RBAC) and
integration on-prem Active Directory (AD) are Also, there are alerting services that help to notify
provided by all platforms. Centralized the support teams and business users.
governance and control can be established.
Reporting
Security Reporting is another core part of an
Security is a core requirement for an automation solution. Business level
automation solution as it deals with metrics reporting can be achieved
sensitive data. Security is two-fold by analyzing transaction level data.
on the cloud. First, cloud hyperscalers Live dashboards with these metrics
ensure the security of the data centers, Cost can offer the business the value
hardware, system software updates, etc. achieved out of automation.
Reduced cost as a benefit of automation is one of
The second part of the security the core goals. Understanding the various pieces of Any visualization tool can be connected
responsibility lies with the users to expenditure is important to know the overall benefit. to the databases on the cloud to create
ensure network traffic and firewalls are Cost governance in cloud is achieved with features dashboards and reports.
configured securely. Enforcing encryption like creating logical entities like org and associating
of data at flight and rest is also a users with these orgs, budgeting provisions and
possibility. Cloud hyperscalers also offer enforced access controls to limit non-relevant
best practices in well-architected usage. Forecasting and Optimization tools help in
framework guides that encompass better financial management.
security, performance,
cost-effectiveness, etc.

8.
Why is Process Automation with Cloud-native
Services Futuristic?
The ever-growing need for digitization has put automation on the fast track, and the need to see
higher ROI is a constant push to innovate. We see that automation tools combine many modules for
which we can find alternatives in the open-source and in the cloud services world.

The explosive growth of the cloud is seen in every industry. Cloud’s benefits of scale, cost, availability,
security and a widespread community are all compelling. So, it becomes natural to bring innovation
in automation by putting together open-source technologies and deploying them on cloud-native
infrastructure leveraging the managed services by the cloud hyperscalers.

6 |
 Figure 1: Traditional vs. cloud expenditure 2019-2025 | Source - Gartner® Press Release, Gartner Says More
Than Half of Enterprise IT Spending in Key Market Segments Will Shift to the Cloud by 2025, February 9, 2022.
https://www.gartner.com/en/newsroom/press-releases/2022-02-09-gartner-says-more-than-half-of-enterprise-it-
spending. GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and
internationally and is used herein with permission. All rights reserved.

Some of the reasons for the exponential growth of automation in cloud-native services can be
attributed to the following:

Serverless Computing Automation benefits from ease of deployment

Serverless computing is computing performed and delivery of bots with containers.Cloud
on managed infrastructure. It removes the hyperscalers also offer a serverless option to
responsibility for back-ends from IT and run containers, thereby giving all benefits of
DevOps teams and enables engineers to focus serverless computing.
solely on application or code functionality.
As per our experience, serverless computing Cloud Security
helps automation benefits in terms of higher Cloud security is a set of policies, technologies
ROI due to the reduced effort and cost on and applications to protect cloud computing
infrastructure (no dedicated servers) and gains environments. Major cloud providers have
from the flexibility to rapidly scale up or down included enterprise-grade security features in
based on demand. their services. We understand that automation
requires the highest security as it mainly deals
Cloud-native Services - Containers with sensitive confidential data. The cloud
Containers are packaged applications, hyperscalers provide the physical security of
operating systems and data. The use of the data centers along with varying levels of
containers has been essential to expanding authentication and authorization options to
cloud services and is the base on which ensure the highest level of security. It includes
cloud-native applications are built. Containers but is not limited to private networks within the
give the automation teams the option to pick cloud (VPC – Virtual Private Cloud), network
and choose various combinations of coding security, data encryption at rest and in transit,
languages, dependencies and platforms. RBAC, resource level access restrictions, etc.

| 7
Managed Services Automation ROI increases if spending is limited
Cloud hyperscalers provide a spectrum of to the solution and infrastructure is paid per
managed services from databases, workflow usage. It eases the burden on the infrastructure
orchestrators, queues and notification services team, where they are not required to plan and
to monitoring and logging. Automation benefits procure hardware ahead of its desired usage
from these services in integration, workflow, period. Cost is also saved on maintenance.
monitoring, etc. These services provide the
backbone of the automation solution and offer Infrastructure as a Code
cost and maintenance advantages, as with the Infrastructure as Code (IaC) manages and
other case in point here. provides cloud infrastructure through code.
Automation benefits from this as all the required
Growth of Hybrid Cloud infrastructure is provisioned with code, which
Pairing private, on-premises data centers with is versioned and saved on source code
public clouds has allowed companies to keep repositories, with no support from the IT team
the critical systems/data on-prem and move the while achieving total DevOps. This provides
rest to the cloud. Hyperscalers provide a private rapid, instant, anytime deployment with
link option to connect VPCs, cloud services and no errors or support.
on-premises networks without exposing the
High-Performance Computing (HPC)
traffic to the public internet. Automation benefits
significantly with an expanded scope due to the All major cloud providers offer HPC instances,
hybrid cloud. Automation solutions running on which are highly parallelized and offer huge
the cloud can securely process both the computing capacity at an affordable cost.
on-cloud and on-prem applications. Automation benefits from HPC systems used
for compute-intensive processing and building
Shift from Capex to OpEx Artificial Intelligence/Machine Learning models.
Users can now save significant amounts of AI/ML models can be integrated with the
money on hardware and data centers and automation flows resulting in intelligent
instead buy services on an as-needed basis. automation.

9.
Use Case
A large wealth management firm in the US achieves high efficiency and scalability while reducing
costs by automating processes with cloud-native services.

Business Challenge
The client’s existing Dormant Account Processing was time-consuming and highly manual. It involved
verifying accounts against the mainframe system, validating data against financial intermediary
websites, performing a manual QC of the account data with the business SME and then creating
accounts in the mainframe system. They aimed to automate their processes to bring efficiency and
scalability and free up the key resources to focus on critical work while saving costs.
Types of Automation: Web Automation, API processing.

8 |
Solution
To enable bot for the Dormant Account Processing, we used AWS cloud, leveraging open-source
libraries and managed cloud-native services. To automate the infrastructure provisioning,
we built a DevOps pipeline using SAM templates and Cloud Formation Stacks from AWS.
We delivered the following services in an integrated manner:

• The mainframe system operations had a • Leveraged IAM (Identity and Access Management)
wrapper exposing them as APIs. These APIs policies to configure the least privilege policies
allowed the automation bot to perform CRUD for resources and Role-Based Access Control
operations for each screen in the mainframe (RBAC) for the users.
system. API processing was done using
• Deployed VPC, subnets and security groups
Java code and deployed on the serverless
to ensure maximum security for the solution
platform, AWS Lambda.
on the cloud.
• Web automation was done using open-source
• Encryption of the data at rest and in flight
Selenium library containerized using docker
was done to safeguard data in the database,
and deployed to EKS Fargate (serverless)
s3 buckets and while being moved between
docker orchestration service. Further, scaling
various services. Leveraged CloudWatch
was done based on the volume of the input.
logs, metrics and alarms to help log and
• Utilized managed workflow orchestrators - monitor the solution.
AWS Step Functions and Event Bus -
• Common dependencies were built into a
AWS EventBridge; serverless databases –
maven project and added as a Lambda layer.
AWS RDS. Workflow QC was done by the
business SME and the handshake between • Spring Cloud was used as the framework to
the bot and the SME was done through put together the automation code in Java.
emails using the AWS SES service.

Benefits
• Automated 5 FTEs work
• Built on open-source technologies with zero licensing cost
• Eliminated manual work for infra setup
• No upfront capacity provisioning because of managed infrastructure from the cloud hyperscalers
• Enabled role-based access to provide required-only access
• All infrastructure provisioned was pay-per-use
• Always available automation on the cloud,
practically zero downtime
• Secured data through encryption
• Constant monitoring and alerting
to help support teams

| 9
10.
What is the Difference Between
Typical RPA Tools vs. Automation Using Cloud?
For automation using the cloud, the following needs to be considered:

Development Additional Responsibilities

Automation is implemented using open-source Although infrastructure is set up as a
libraries like Selenium for web automation, one-time configuration to be repeatedly used,
Apache POI libraries for excel-based work, etc. the responsibility to develop the code to
provision them is with the development team.
Detailed Design
Design requires understanding the features and Scale
limitations of the various services being used. Scaling up and down the capacity is an
Cloud Resource Access policies are defined at automated task with the cloud.
the basic task level like read, create, update,
etc., and must be detailed during the design Documentation
phase. Region-specific restrictions for services, Extensive documentation is available on various
with default quotas, must be understood. cloud services, with details on each feature and
examples of how to use them. This aids the
Infrastructure design, configuration and development.
No planning is necessary for capacity
or maintenance with this new approach. Huge Community
Provisioning and basic setup of the The cloud community is vast, and most tech
new servers are also avoided here. issues that are encountered are highly likely to
have been addressed already.
Infrastructure as Code
Infrastructure is provisioned using configuration
templates, unlike RPA tools, where this is manual.

11.
Conclusion
Process automation using the cloud will become the next big thing shortly. We can see from the
above that building automation solution on the cloud provides huge cost, effort and scale benefits.

We believe it is a natural step to build your next automation solution open-source and on the cloud.
Having gone through the journey of building and delivering on this technology stack is a proof of how
production-grade automation solutions can be achieved with this stack. Despite its challenges related
to pro-code/intensive coding requirements, the overwhelming benefits of this novel approach make
it a compelling option to consider for automation solutions. With the rapid innovation and adoption of
the cloud, such an approach makes sense not only for the ‘now,’ but also for the future.
10 |
Authors
Vinod Kumar Mohan
Automation Consultant, Senior Manager

Veena S
Enterprise Automation Leader, Vice President

Renjeev Kolanchery
Enterprise Automation Head, Senior Vice President

About Mphasis
Mphasis’ purpose is to be the “Driver in the Driverless Car” for Global Enterprises by applying next-generation design, architecture and
engineering services, to deliver scalable and sustainable software and technology solutions. Customer centricity is foundational to Mphasis,
and is reflected in the Mphasis’ Front2Back™ Transformation approach. Front2Back™ uses the exponential power of cloud and cognitive to provide
hyper-personalized (C = X2C™² = 1) digital experience to clients and their end customers. Mphasis’ Service Transformation approach helps ‘shrink
the core’ through the application of digital technologies across legacy environments within an enterprise, enabling businesses to stay ahead in a
changing world. Mphasis’ core reference architectures and tools, speed and innovation with domain expertise and specialization, combined with an
integrated sustainability and purpose-led approach across its operations and solutions are key to building strong relationships with marquee clients.
Click here to know more. (BSE: 526299; NSE: MPHASIS)

For more information, contact: marketinginfo.m@mphasis.com

USA UK INDIA
Mphasis Corporation Mphasis UK Limited Mphasis Limited
NR 07/10/22 US LETTER BASIL7605

41 Madison Avenue 1 Ropemaker Street, London Bagmane World Technology Center

35th Floor, New York EC2Y 9HT, United Kingdom Marathahalli Ring Road
New York 10010, USA T : +44 020 7153 1327 Doddanakundhi Village, Mahadevapura
Tel: +1 (212) 686 6655 Bangalore 560 048, India
Tel.: +91 80 3352 5000
Copyright © Mphasis Corporation. All rights reserved.
2 |
8 | |11
www.mphasis.com5
7