Enterprise Strategy

Management Issue
July 14, 2016

Understanding Robotic Process Automation, Part 1:

Maturity, Adoption and Future
By Erik Dorr, Martijn Geerling and Jim O’Connor

Executive Summary
Robotic process automation has potential for opportunistic, business-led deployments and high ROI, especially for
global business services organizations. However, RPA comes with several risks and caveats. First, while business cases
may be very compelling, The Hackett Group considers RPA to be tactical rather than transformative in nature. Second,
companies must recognize that the lifespan of RPA in non-industry-specific business processes is determined by the
lifecycle of the legacy applications supporting these processes. How long these legacy applications will be retained will
vary widely between companies. This should not prevent organizations from exploring RPA, but underscores the need
for deliberate planning and establishment of strong governance throughout the entire RPA lifecycle.

About this research Understanding Robotic Process Automation

This is the first in a three-
“Robotics” is making a rapid entry onto the agenda of business services organizations.
part series on robotic
process automation. While most of the actual deployments of robotics to date have been in industry-specific
In Part 1, we define RPA,
processes (e.g. financial services, utilities, and telecom), successful applications in non-
analyze its current-state industry-specific (e.g. HR, procurement, IT and finance) processes are emerging. These
maturity and adoption, and business functions are scrambling to gain a better understanding of the technology
project its future adoption
and its potential to deliver changes in efficiency performance.
trends.
In Part 2, we analyze RPA’s
value proposition and look The term “robot” is traditionally applied to machines executing human-like tasks. More
at how RPA is used. recently, the term has evolved to describe a class of software applications – software
In Part 3, we present a robots1 – that automate work by emulating tasks executed by human operators. This
model of RPA use-case form of automation is referred to as robotic process automation, or RPA.
evaluation criteria and
offer recommendations
for developing business RPA is usually defined narrowly, referring to technology that automates repetitive, rule-
cases and a roadmap for based, routine tasks (see sidebar on following page). At present, rule-based robotics
adoption. technology has the highest potential for opportunistic, short-term deployment in
Parts of this series draw traditional back-office processes, which is the focus of this research.
on a roundtable featur-
ing members of our RPA vendors position their product as distinctly different from traditional types of
Global Business Services
technology, emphasizing its capacity to replace human operators and using terms
Executive Advisory Pro-
gram whose companies like “digital workforce.” However, from a technical perspective, RPA is just a type of
are early adopters of software application, i.e., technology. In simple terms, a robot is a software program
RPA technology in G&A that runs on an end user’s computing device (e.g., a PC or laptop) emulating tasks
processes. The event was executed by human operators. Defined more precisely, RPA is a method to integrate or
conducted in early 2016.
automate the execution of repetitive, rule-based tasks or activities through deployment
Additionally, to support of software running on desktops and acting as an agent for a human operator.2
this research we inter-
viewed and conducted
briefings with the major
RPA technology vendors
and a number of BPOs.
1
Adding some confusion to the RPA lexicon, the term “bot” ( which predated the term software robot) has tra-
ditionally been used to describe the same technology as a software robots. Bot and robot designate the same
technology and can be used interchangeably.
2
Cognitive computing is excluded from our definition of RPA and is not addressed in this research.

© 2016 The Hackett Group, Inc.; All Rights Reserved. | ES0070 Enterprise Strategy I The Hackett Group I 1
RPA and cognitive computing RPA is non-intrusive and implemented through deployment of software robots on end
Technology automating human like tasks users’ computing devices, operating in a managed systems environment. The robots
that require some level of intelligence interact with the user interface (presentation) layer of the business applications in which
and “interpretation” of information (e.g., execution of the tasks or activities is implemented. (Explanations of the various elements
Apple’s SIRI, emerging “chatbot” tech- of this definition of RPA are included in the sidebar beginning on the next page.)
nology) are sometimes also referred to
as software robotics. For example, IPsoft Current State of Maturity and Adoption
refers to its cognitive technology “Ame- The high level of media attention being given to RPA does not necessarily indicate high
lia” as a software robot, whereas IBM is adoption levels. Nor does it say anything about the maturity of the market or technology.
referring to its Watson technology as a • Dominant applications: The most current RPA implementations are found in industry-
cognitive computing platform. However, specific processes such as claims processing in insurance, risk management in financial
IBM does use the term bot to describe services, and customer service in utilities. These processes and associated tasks
applications developed using the Watson are usually high-volume, structured, repetitive and implemented on old – sometimes
platform. This type of technology is more mainframe-based – technology. Typically, the processes are extremely stable. There
commonly called cognitive computing. is no technology migration or modernization roadmap for the systems environment,
and IT-led integration would be difficult and expensive. These processes are eminently
Another term used in the context of suitable for RPA.3
software robotics is artificial intelligence
(AI). AI is the broad discipline of comput- At present, the leading non-industry-specific 4 RPA application is the financial
ing that emulates human intelligence, close and consolidation process. In fact, a number of technology vendors offer
including rational decision making and RPA technology to automate repetitive, rule-based accounting tasks (e.g., account
machine learning. Cognitive computing reconciliations). Although this is the leading back-office RPA application, its adoption,
refers to a collection of technologies while growing, remains very low. Another focus area of RPA in generic business
which has emerged from AI. Example processes is purchase-to-pay.
of cognitive technologies are machine • Global business services adoption: Mature RPA implementations in GBS organiza-
learning, speech recognition and com- tions are most often found in the industry-specific processes referred to above when
puter vision. run in a GBS. Beyond this, some large, mature GBS organizations are in early adop-
tion or the piloting stage of RPA (Fig. 1). Most of the RPA-related inquiries we receive
The three stages of technology- from our Advisory Program members originate from GBS organizations. Our research
enabled execution of work indicates that, while exploratory activity level is very high, few GBS organizations have
Process automation based on business gained insight into full lifecycle costs and benefits of RPA deployments because the
application suites (Stage 1), process inte- technology is relatively new.
gration using business process manage-
• RPA market/technology vendors: The RPA market is in its infancy, and vendors are
ment technology or RPA (Stage 2), and
small in terms of revenue, staffing levels and client base (see sidebar on page 6).
ultimately, “service orchestration” using
These vendors are in the early stages of developing partnerships with service provid-
cognitive computing (Stage 3) represent
ers and implementation partners. They are presently supporting most implementa-
progressively higher levels of technology
tions of their technology themselves, creating capacity bottlenecks. For the market to
enablement of the execution of work.
mature and grow, technology vendors must develop partner channels and new vendors
must enter the market to drive continued innovation and expansion.
While the transition from Stage 1 to
Stage 2 (the focus of this research) offers • Consulting firms/service providers: While some smaller service providers have
dramatic productivity improvement, the worked closely with RPA technology vendors to support implementations, the larger
transition from Stage 2 to Stage 3 will be ones are developing RPA-based commercial offerings more conservatively. Many large
truly transformative in nature and change consulting firms also provide BPO services based on a labor-intensive service delivery
what it means to execute work. model. However, in response to growing demand and projected expansion of RPA
technology, development and implementation services, service providers are starting
Dominant Dominant develop expertise and consulting practices, and establishing partnerships with technol-
enabling human
technology role ogy providers.

Stage 3: Cognitive Decision

Service computing making
orchestration

3
Stage 2: Business Exception While RPA integrates applications at the user interface layer, IT-led integration uses APIs, database access,
Process process handling application integration tools, middleware, workflow or business process management software.
integration manage-
ment, RPA In this context, non-industry-specific business processes are defined as generic back-office processes in HR,
4

Stage 1: Business Clerical procurement, finance and IT. Processes that, while designed and implemented with unique, industry-specific
Process application tasks, data characteristics, but found in all or most industries (e.g., purchase-to-pay) are also considered non-industry-
automation suites entry specific.

© 2016 The Hackett Group, Inc.; All Rights Reserved. | ES0070 Enterprise Strategy I The Hackett Group I 2
 Robotic process automation:
Key terms and concepts
• Automation or integration of applications: While RPA is commonly referred to as
process “automation” technology, a more accurate description in most deployments
would be “application integration” technology. A robot typically replaces a human
operator who manually integrates applications, for example by logging in to multiple
applications and entering the same digital input into these applications, or by taking
the output of one application and entering it into another application. Executed by a
human, this type of integration is sometimes called “swivel-chair integration.”
• Tasks and activities: While the term RPA suggests process automation, in reality RPA
operates at the task and activity level. End-to-end processes consist of a series of pro-
cesses which are decomposed into a set of subprocesses. Subprocesses consist of
activities, which are in turn comprised of a discrete tasks (Fig. A). RPA operates at the
activity or task level. Most business processes require multiple software applications
to execute different activities and tasks. For example, in an end-to-end purchase-to-pay
process, the payment portion of the process may require an approval activity, which
consists of multiple tasks, like extracting accounts payable information from an ERP
system, parsing out approver information in a spreadsheet, sending emails to approv-
ers, obtaining approval, updating records in the ERP, and authorizing payment. In this
example, the payment approval activity requires three applications: ERP, spreadsheet
and email. RPA operates at the activity and task level, and typically across multiple
applications.

FIG. A Process hierarchy

END-TO-END PROCESS Process Process Process Process Process

Sub- Sub- Sub- Sub- Sub-

PROCESSES process process process process process

SUBPROCESSES Activity Activity Activity Activity

ACTIVITIES Task Task Task Task

STEPS, SYSTEM INTERACTIONS System System System System

Source: The Hackett Group

© 2016 The Hackett Group, Inc.; All Rights Reserved. | ES0070 Enterprise Strategy I The Hackett Group I 3
 • Repetitive and rule-based: RPA executes activities and tasks which are repetitive and
follow a fixed set of rules, which may include conditional logic. In the payment approval
example above, the robot is programmed to apply objective business rules (i.e., not
subject to judgment or ambiguity), such as the threshold for human authorization of
the payment. This logic includes rules for exception handling.
• Non-intrusive: RPA does not require development of code, nor any form of direct
access to the code or database of the applications integrated with software robots.5
This is a critical feature because it allows the business to develop robots without the
involvement of the IT organization.
• Managed systems environment: Software robots are not developed or deployed on
desktops in isolation. Design, development and deployment are managed in a central-
ized, controlled environment. One of the reasons so many organizations have become
interested in RPA is that the technology has matured to the point that robust, enterprise-
class deployment is feasible. While in principle robots can be developed without any IT
involvement, governance and provision of the managed systems environment requires
collaboration between the business and the IT organization.
• User interface layer: Robots emulate steps (i.e., mouse clicks and keyboard strokes)
executed by human operators on desktops, interacting with the user interface layer
of the integrated applications. This approach dramatically reduces the complexity and
effort involved in developing robots and minimizes dependency on IT resources.

5
Direct access of the underlying application would typically take place through an application programming
interface (API), point-to-point integration or direct access of the application’s database, and require involvement
of developers in the IT organization.

© 2016 The Hackett Group, Inc.; All Rights Reserved. | ES0070 Enterprise Strategy I The Hackett Group I 4
 FIG. 1 Investment and planned investment in RPA solutions

By function By process group

6% GL accounting 7% 26% 67%

Cash
Finance 11% 82% 7% 26% 67%
disbursement
Revenue cycle 13% 9% 78%

1%
6%
Payroll 12% 88%
Human
93%
resources Staffing 12% 88%
services

Service desk 25% 75%

Information End-user
technology 15% 85% 17% 83%
device support
Network
18% 82%
management

INVESTED PLAN TO INVEST NO INVESTMENT PLANNED

Source: Global Business Services Performance Study, The Hackett Group, 2016

• Business process outsourcers: Traditional BPO business models are based on realiza-
tion of labor arbitrage. It is too early to predict RPA’s impact on the BPO market with
a high level of certainty. However, the technology does have the potential to disrupt
the industry by changing its economics. In this scenario, large pools of offshore labor
doing repetitive, rule-based work would be made obsolete. Incumbent BPOs may try
to defend their traditional business model by resisting RPA adoption. Alternatively, they
may develop their own RPA capabilities, working with existing clients to migrate from
a labor-intensive to a more technology-intensive service delivery model. While some
BPOs are developing their own RPA technology, most are partnering with third-party
technology vendors. Additionally, new BPOs are entering the market, positioning them-
selves as “native” RPA-based service providers. Their plan is to leapfrog past incum-
bents in terms of efficiency and quality of service delivery.

While outside the scope of this research, it is worth noting that leading BPOs are
making strategic bets on broader digital transformation, which involves fundamental
redesign of work. This involves much more than inserting RPA into existing business
processes, and is predicated on the maturation of cognitive technologies. While RPA is
tactical in nature, BPOs’ ability to support broader digital transformation will be strategic
and a competitive differentiator.

The Future of RPA

RPA is a poorly understood, emerging technology with vast potential for opportunistic,
business-led deployment. However, major supply-side constraints exist which organiza-
tions must understand in order to plan their RPA adoption course. While projections of
any new technology are fraught with uncertainty, we offer the following outlook on the
future of RPA:
• RPA is poised for imminent takeoff and rapid growth over the next several years.
Rising awareness, maturing technology and pioneers who are already climbing the
learning curve are setting the stage for wider adoption. Growth will not be limited by
demand-side factors, because processes for which RPA can be deployed economically
are abundant. Rather, adoption will be held back by a shortage of development and
implementation resources, plus a dearth of other RPA-related competencies required
for successful adoption. With ample deployment opportunities, growth will be a
function of the pace at which technology, service providers and user organizations can
build their capacity and organizational competencies to implement and support RPA.

© 2016 The Hackett Group, Inc.; All Rights Reserved. | ES0070 Enterprise Strategy I The Hackett Group I 5
 • Industry-specific applications will continue to attract most of the attention in RPA
Leading general-purpose
for the next two or three years. Operating under resource constraints, RPA deploy-
robotics vendors
ments will gravitate toward the opportunities with the greatest potential return. For
The leading vendor of general-purpose
the next several years, these will continue to be industry-specific processes. Generic
RPA technology, Blue Prism, went
RPA applications for business services will remain limited to specific processes such
public in Q1 of 2016. It reported
as purchase-to-pay, financial close and consolidation, and other accounting processes.
revenue of £6.0 million (US$8.8
Further, because business applications supporting such processes are being modern-
million) and growth of 35% in 2015,
ized and consolidated, RPA lifecycles will be shorter, offering a much less compelling
with a total of 74 customers. The two
business case than that for industry-specific processes.
other leading general-purpose RPA
vendors, UI Path and Automation • RPA will be evolutionary, not revolutionary. RPA will not revolutionize delivery of
Anywhere, are substantially smaller business services, but rather offer (potentially large) evolutionary change in specific
than Blue Prism. By any standard, process domains. Over the longer term, it is cognitive technologies that will revolu-
these are very small companies in the tionize service delivery. Unlike RPA, these are truly innovative in nature, allowing the
technology industry. complete redesign of work, including automation of non-deterministic processes and
knowledge work.
• RPA adoption will be heavily concentrated in GBS organizations and BPOs.
Process variance is the enemy of RPA. The tasks most suitable for RPA are high-
volume, stable, standardized, repetitive and rule-based. These are precisely the types
of processes companies are running in their GBS organizations or that are slated for
eventual migration to the GBS. Process rationalization, standardization and upscaling
are typically done in a centralized GBS environment. Therefore, RPA is a natural
progression of performance improvement realized through movement of work into a
GBS. Furthermore, the management, maintenance and governance of RPA require
business-IT collaboration in a centralized model. This also heavily favors the GBS as
the service delivery organization of choice for RPA.
In addition to captive GBS organizations, the main impact of RPA will be felt in BPOs.
As discussed earlier, many of the processes and associated tasks suitable for RPA
are executed by BPOs, and RPA has the potential to fundamentally change the
economics of delivery of this type of work. BPOs must either embrace RPA and help
customers harvest the gains, or risk losing business and be disrupted by native RPA-
based competitors.
• RPA in non-industry-specific business processes is a tactical technology. The
compelling business cases for RPA deployments in industry-specific applications will
drive significant investment and generate solid financial returns. However, service
delivery organizations should plan RPA for non-industry-specific business processes
more conservatively. The lifecycle of RPA is tied to those of the underlying business
applications that are integrated though RPA. Consequently, the RPA lifecycle is a
function of the rate of modernization and consolidation of the business applications
landscape, which will vary by company. Service delivery organizations should adopt
RPA as a tactical technology and plan its lifecycle based on the broader business
application roadmap. Failure to sunset RPA technology and modernize application
platforms may lock companies into legacy technology that will hamper their agility and
limit their ability to respond to business demands. Conversely, RPA permits compa-
nies to defer major investments in application upgrades and avoid associated risks.
Industry-specific RPA applications are likely to have much greater longevity since no
clear technology migration path exists for many of the custom-built or heavily custom-
ized systems integrated through RPA.
• Poorly planned RPA may slow down process redesign and application
modernization. While in and of itself RPA is a technology innovation, paradoxically,
if planned poorly, it may slow down innovation and broader digital transformation. By
creating a user-interface-level integration layer on top of legacy systems, organizations
may actually be locking themselves into their legacy processes and systems. While
this consideration should not prevent them from undertaking economically sound RPA
investments, they should be aware of this risk and plan their technology modernization
roadmap accordingly.

© 2016 The Hackett Group, Inc.; All Rights Reserved. | ES0070 Enterprise Strategy I The Hackett Group I 6
Strategic Implications The Hackett Group (NASDAQ: HCKT) is
RPA offers opportunities to improve delivery performance of repetitive, rule-based, an intellectual property-based strategic
stable and high-volume business processes, primarily by eliminating human costs and consultancy and leading enterprise
errors in execution. However, service delivery organizations need to exercise caution benchmarking and best practices
when selecting use cases, committing resources or selecting technologies and service implementation firm to global companies.
providers. Deliberate planning – including rigid modeling of cost and benefits as the Services include business transformation,
basis for business cases – is critical. enterprise performance management,
working capital management, and global
Related Hackett Research business services. The Hackett Group also
“Understanding Robotic Process Automation, Part 2: Value Proposition, Deployment provides dedicated expertise in business
Model and Use Cases,” July 2016 strategy, operations, finance, human
capital management, strategic sourcing,
“Understanding Robotic Process Automation, Part 3: Developing Business Cases and procurement, and information technology,
Getting Started,” July 2016 including its award-winning Oracle EPM and
SAP practices.
About the Advisors
The Hackett Group has completed more than
Erik Dorr 11,000 benchmarking studies with major
Vice President, Strategic Research, and Lead Author corporations and government agencies,
including 93% of the Dow Jones Industrials,
Mr. Dorr has over 20 years of experience in consulting, research and 86% of the Fortune 100, 87% of the DAX 30
advisory roles in information technology strategy, enterprise application and 51% of the FTSE 100. These studies
suites and business process reengineering. Before being named to his drive its Best Practice Intelligence Center™,
current position, he was Senior Enterprise Research Director. Prior to which includes the firm’s benchmarking
joining The Hackett Group, he held a number of senior management metrics, best practices repository, and best
positions, including Vice President of IT at a global manufacturing practice configuration guides and process
company, where he was also a member of the executive leadership team. flows. It is this intellectual capital that enables
The Hackett Group’s clients and partners to
achieve world-class performance.

Martijn Geerling
Practice Leader, Global Business Services Advisory Program

Mr. Geerling has over 15 years of consulting experience in strategy

development, business process redesign and implementing sourcing
strategy, both shared services and outsourcing. During this time he
has worked with business services organizations of leading global Email: [email protected]
companies across various industries. Prior to joining The Hackett www.thehackettgroup.com
Group, he worked at KPMG Consulting assisting clients in Europe and
Asia with finance-function transformation, business process redesign, and risk and Atlanta +1 770 225 3600
compliance management. London +44 20 7398 9100
Sydney +61 2 9299 8830

Jim O’Connor Atlanta, Chicago, Frankfurt, Hyderabad,

Global Practice Leader, Global Business Services and Finance Advisory Programs London, Miami, Montevideo, New York,
Paris, Philadelphia, San Francisco, Sydney,
Mr. O’Connor has over 20 years of both industry and consulting Vancouver
experience focusing on finance transformation. He has particular
expertise in strategy and organization design, business process
design, strategic cost reduction, reporting, planning and performance
management, BI and financial systems, shared services and outsourcing.
In these roles, he has advised client executives in a wide range
of industries including consumer products, financial services, higher education,
manufacturing, retail, and utilities. Previously, he led the CFO Services practice at North
Highland, a global consulting firm, and before that, focused on finance transformation
and strategy at Archstone Consulting, now a part of The Hackett Group.

This publication has been prepared for general guidance on the matters addressed herein. It does not constitute professional advice.
You should not act upon the information contained in this publication without obtaining specific professional advice.

© 2016 The Hackett Group, Inc.; All Rights Reserved. | ES0070