Beyond the hype

__

Moving
ahead
with
Intelligent
Automation
The advances of artificial intelligence
are astounding

UBS bank has explored the potential of AI to create of the huge amount of financial and economical data
the best possible customer experience. No wealth available.
management advisor is able to know all the latest
information on every asset category and every market. The underlying IBM Watson technology provides the
For a knowledgeable and unbiased advice, FaceMe1 backend intelligence2 to the avatar. This example
modeled an avatar after the Chief Investment Officer from the Banking Industry illustrates how advances
in Switzerland. This avatar is the second expert in the in AI will bring the customer interaction to completely
room with the client. It is trained to look for important new levels, finally supported by zero-touch digital
information relevant for an investment decision out processes from the front-office to the back-office.

1) https://www.ibm.com/cloud/blog/ubs-brings-financial-expertise-to-customers-using-innovative-digital-assistants
2) https://www.ibm.com/case-studies/faceme-cloud-ai-digital-humans-customer-experience

© faceme

2
Beyond the hype

Robotics Process Automa- Utopia

Unprecedented levels of productivity gains will be
tion combined with Artificial driven by “automation arbitrage” where simple
and not so simple tasks performed by humans are
Intelligence (AI) and Block- supported or replaced by technology that tirelessly
performs 24x7
chain drive the next wave of
business transformation Dystopia
Humans will perform less than half of all workplace
tasks in 2015, compared to 71% in 2018, according
to the World Economic Forum’s “Future of Jobs
But the hype has reached the 2018” report. Hundreds of millions of jobs will get
lost world-wide
peak of inflated expectations

“Humans are underrated”

Elon Musk, Tesla

Analysts agree that 2019 the hype around automation technologies will pass beyond the climax
and we enter the trough of disillusionment. The hype is underpinned by great success stories,
the disillusionment comes with the experience that there is no magic in the technology itself.
Elon Musk admitted that excessive automation was a mistake and also Toyota, still a relentless
pioneer in robotics, increased the human part in automotive production after facing quality issues
and falling profits due to high levels of automation

So – what’s in for you in 2019 and beyond to reap the benefits of this new wave of automation
and steer clear of the pitfalls?

3
Understanding the opportunity

Intelligent automation is changing the way enterprises done, bringing new operating models, organizational
operate by using advances in technology to optimize constructs and value creation opportunities.
processes, personalize customer experiences and
enhance decision making: The value of automation primarily comes from the ef-
ficiencies it creates by taking over simple but tedious,
Robotics process automation (RPA) is automating re- repetitive tasks from work packages that were cove-
petitive manual tasks, Artificial Intelligence (AI) is ex- red by humans before. The impetus is not necessarily
tending the automation capabilities through learning, to replace people, but to transform the work that hu-
reasoning, decisioning and discovery capabilities and mans do and create new ways of working to serve the
Blockchain is bringing in trust across ecosystems. clients best. We call this the Cognitive Enterprise.

Automation will impact every part of the enterprise, Cognitive Enterprise:

and it will permanently change the way work gets https://ibm.co/cognitive-enterprise

4
Beyond simple automation

Basic automation is good for rules-based, repetitive Example: A Fortune 75 global consumer goods com-
tasks with well-structured activities, clearly defined pany used advanced automation and resolved work-
rules taken from well-structured data sources, and flow problems upwards of 30 percent more quickly
systems that result in visible and measurable outco- and improved employee productivity by upwards of 50
mes. percent

An example would be the updating of an order in the

order system which would require several checks like
whether the additional goods are on stock and if the
larger order could still be delivered on the previously
planned date. A human would follow a sequence of
steps involving the log-in in various systems, making
lookups and updates. Robotics process automation
has been hugely successful at automating such kind of
processes.

Obviously, this transformed process might be re-de-

signed such that it becomes a self-service process for
a customer, available 24x7 provided that the subse-
quent business processes can deal with the informa-
tion produced via the self service interface.

Simple automation of processes can eliminate errors,

reduce biases and perform transactional work in a
fraction of the time it takes humans.

Basic Robotics Process Automation has demonstrated

up to 75 percent cost savings on repetitive tasks com-
pared to human performance, with 25 to 50 percent
being the generally reported outcome.

5
 Beyond simple automation

However, many processes aren‘t that simple, or the Possibly not all required information was found or
simple part of it only covers part of the volume. It contained in the document. But business rules could
might well be that only a tiny fraction of customers apply to deduce the order number e.g. based on who
would engage over the self-service interface, while sent the order and whether there is a unique current
others might stick to sending e-mails, letters, possibly order that contains the item for which the quantity has
even fax. to be changed.

Augmenting the human part of reading the request Finally, the interaction between the human and the
and extracting the relevant information – like: what backend system could be transformed with the help
is the order number, who is the sender, what item is of an intelligent assistant with whom the expert or
to change – is something that could be done with customer could have a dialogue to get this job done,
Artificial Intelligence combined with technology to and we will shortly see an example for this.
make rule-based decisions.

The AI would be trained to extract and understand

the relevant information from the text with Natural
Language Processing and classification. If the AI
system has high confidence in the results the process
might continue automatically. Otherwise the human
might be presented an extracted list of what the
system found, side by side with the original text for
validation.

AI is a key enabler of intelligent automation

Operations After Sales

Pre-sales Sales

• Virtual Assistants • Optical Character Recognition • Machine Learning • Robotic Process Automation
• Chatbot, Messaging • Natural Language Processing • Deep Learning • Microservices Architecture

6
Steer clear of disillusionment

So – how could you reap the benefits of this new wave Similarly one needs to take into account how repeti-
of automation and steer clear of the pitfalls? Setting a tive or unique the tasks are. The less repetitive and
clear course and realistic expectations should get you simple, the less automated you can get. But that’s also
started or help you pull through. the good news: “humans are underrated”. Experts are
not particularly fond of the repetitive and stupid parts
Early successes by applying simple robotics process of their work. On the contrary, they will appreciate the
automation might create unreasonable expectations. aid of an AI-powered assistant that munges through
Automation comes at a cost and pays off depending huge amounts of data to come back with the reduced
on the volume. Intelligent automation even more so list of relevant pieces of information on which to conti-
as you need to have a sufficient amount of training nue their work of reasoning, making decisions, impro-
data and also put your specific training efforts into the ving customer experiences – in tasks as varied as legal
development. contract reviews, medical treatment decisions, claims
analysis, fraud management, cyber threat analysis or
supply chain optimization.

Uniqueness of work
Unique

ADMINISTRATIVE WORK EXPERT WORK

Increase the
productivity
Assist humans in their of employees
decision making, with with Cognitive
next-generation workflows Systems
e.g. in customer service

Reduce manual effort,

minimize the burden of
compliance and audit

Reduce departmental
work by combining nextgen
workflow design + low code
apps
Supplant human-labor
with ‚bot’-driven tasks
Repetitive

REPETITIVE WORK DEPARTMENTAL WORK

High volume Low volume

7
Clear course towards The Cognitive
Enterprise

The bigger strategic context of Intelligent

Automation is the evolution to the Cognitive
Enterprise which incorporates exponential Success factors for becoming
technologies as AI, RPA and Blockchain to create
results at scale. a Cognitive Enterprise
The 19th Global C-suite Study by the IBM Institute
for Business Value uncovered the CEO’s top five
drivers to accelerate performance.

Intelligent automation will directly contribute to

1. Focus on your business platform
creating more agile operations by driving the slack of
repetitive, high-volume tasks out of workflows. This
in turn requires to invest in people so they can better
focus on what sets humans apart from machines. As
Mitsuru Kawai, head of manufacturing and executive
vice president of Toyota, put it,
2. Leverage your incumbent data

“Robots cannot improve

processes. Only people can
do that, and that is why they
3. Reinvent your workforce
are always at the center of our
attention.“

Sources: 19th Global C-suite Study, IBM Institute for Business Value, 2018,
Roland Berger, “Automotive manufacturing requires human innovation”, 2018

4. Create intelligent workflows

8
Clear course towards The Cognitive
Enterprise

Re-imagine processes to execute at digital speed to accelerate

performance and enable an agile enterprise

http://ibm.com/Garage

Getting there requires enabled professionals, supported by intelligent

workflows, which are infused by insights derived with data.

Human-centered outcomes require empathy for the people you serve.

Leverage the IBM Garage, our award winning prescriptive Method to re-imagine
your processes by putting the human in the center. The methodology is fueled
by a collaborative culture incorporating:

- IBM Enterprise Design Thinking

- a transformation framework
- development & operations principles (DevOps)
- based on agile, distributed agile
- applying lean
- incorporating talent development & change

9
Case Studies

10
 1
Case Study 1:
Intelligent Mail Assistant
Business problem
The volume of online inquiries is growing 23% a year.
Maintaining the quality of client relationships, while
dealing with an ever-rising stream of customers and
client requests, meant reinventing the role of client
Crédit Mutuel, one of France’s advisor or losing their competitive edge

leading banks, has over 5,000 Solution

branches that receive more than After running a diagnosis of how client advisors
were spending their time, Crédit Mutuel found that a
350,000 online inquiries a day significant part of their work involved answering simple
and repetitive questions. With this in mind, the bank
turned to IBM to find a solution that could speed up
everyday processes and give client advisors time to
address more complicated and nuanced problems.
Watson learned from four virtual assistants and an email
analyzer, analyzing banking and insurance vocabulary
and building 50,000 lines of dialogue. Crédit Mutuel
rolled out solutions progressively, across different
lines of business. Watson has been assisting 20,000
customer advisors strengthen customer relationships
in 5,000 branches. Once Watson was adopted, client
advisors could answer customer questions 60% faster.
Watson continues to be trained by business experts on
thousands of customer inquiries

Watson answers 50% of the

350,000 daily emails automatically
and helps advisors to find the right
answers to the rest 60% faster

https://www.ibm.com/watson/stories/creditmutuel/ 11
 2
Case Study 2:
Replenishment Advisor
Business problem
Dynamically changing market conditions across
continents and time zones create challenges in the
largely distributed supply chain. The goal was to
optimize the replenishment planning and minimize
BASF is one of the world’s largest disruptions

chemicals groups, headquartered Solution

in Ludwigshafen, Germany, The BASF Nutrition & Health Supply Chain team
engaged IBM Services™ to help build a PoC. The goal
with 390 production sites in 90 was to evaluate how cognitive technology based on AI
countries, and. BASF employs and machine learning could be used to build a powerful
Replenishment Advisor tool.
more than 122,000 people and
The Replenishment Advisor tool could be built quickly
achieved annual sales revenues using IBM Cloud™ and IBM Watson AI services, a
of about EUR 63 billion in 2018 robust integration platform that allows expert users
to connect structured and unstructured data from
disparate systems to deliver a unified overview of supply
chain operations.

With the success of the PoC, BASF Nutrition & Health

could demonstrate the potential of cognitive supply
chain management to the entire group.

Together with IBM, the team is now evaluating ways to

bring the PoC into production.

The solution increases visibility

on operation status, enabling
better supply chain replenishment
decisions.

It supports more comprehensive

planning, monitoring and
optimization of a significantly
larger portfolio

https://www.ibm.com/case-studies/basf 12
 3
Case Study 3:
Intelligent automation in insurance
Business problem
ROLAND wanted to explore the potential of Intelligent
Automation with the help of Natural Language
Processing (NLP) and AI to adress two business
challenges. The first is to detect so far unnoticed
The ROLAND Group is a recourse potential, which in the industry is estimated
to be in the order of 3% of claims. Secondly, ROLAND
network of specialists operating wants to improve efficiency and customer loyalty by
throughout Europe and accelerating the handling of coverage requests.

headquartered in Cologne. Solution

ROLAND and IBM Services collaboratively ran a fast-
paced agile project to build a Minimum Viable Prototpye
ROLAND Rechtsschutz- to prove feasibility. Therefore, complex lengthy legal
documents had to be analyzed to detect e.g. the field
Versicherungs-AG is a premium of law or an indicator of recourse potential of closed
provider in the field of legal damage cases. NLP and AI models were tuned until they
met the threshold for a solid business case.
protection Hence both prototype solutions now are in the course of
implementation throughout 2019.

Through intelligent automation

ROLAND expects to substantially
increase efficiency and improve
loyalty.

By identifying additional recourse

potential, ROLAND can increase
the profit

https://www.roland-rechtsschutz.de/unternehmen/unternehmen.html 13
4
Case Study 4:
Sales order automation
Business problem
The client was looking for a way to improve efficiency of
its inside sales agents by reducing time in the creation
of sales offers to free valuable resources from repetitive
simple tasks. Yearly more than 250,000 sales orders
This industrial company is needed to be manually processed to fulfill the requests
of hundreds of customers.
operating worldwide in more
than 70 countries and specialized Solution
IBM designed and developed an AI based sales order
in producing and processing automation solution. A responsive user interface was
technologically sophisticated used to dynamically drag & drop mails from Outlook to
get a sales order recommendation. Once the mail was
high performance materials imported, the customer and multiple steel products
were extracted from the mail using Watson Natural
Language APIs. The information identified was then
used to automatically make material suggestions
searching through customer databases. After the user’s
confirmation of the displayed sales order, existing
interfaces were leveraged to directly import and
complete the order in the customer’s SAP system.

Through the implemented AI

automation solution, the process
efficiency could be increased by
40 %

14
5
Case Study 5:
Sales and Inventory Reconciliation
Business problem
The company sought to improve its sales and inventory
reconciliation processes. These processes consisted of
manual reconciliation of over 1,000 reports per day. The
labor-intensive, repetitive and time-consuming process
This global cement and building created a high risk of defects.

materials manufacturer has Solution

more than 40,000 employees. First, the fit of the process for automation was
evaluated. Then, using IBM’s library of business process
With operations in more than frameworks, the team mapped out an optimized
50 countries, the company reconciliation process. This process was implemented
with Blue Prism.
sought to improve its sales
and inventory reconciliation
processes The automated process helped
the client reap benefits through
a 39% reduction in cycle time a
19% reduction in manual errors
and in addition freed up experts
who can now focus on higher-
value activities

15
6
Case Study 6:
Intelligent Legal Assistant
Business problem
More and more clients try to solve their legal problems
online. But human legal specialists do not scale. 
Lawyers today spend 80% of their time on problems
they or their colleagues have solved before. 
QNC is building platforms for
Solution
online legal advice and legal tech After running a deep analysis of the existing legal
serving hundreds of lawyers, content of more than 200.000 cases QNC has set up
with IBM an intelligent legal assistant based on Watson
in-house legal departments and that is deeply integrated into the lawyer’s workflow
hundreds of thousands clients and suggests answers to repetitive legal problems.
QNC has rolled out the system to different law firms
monthly and in-house legal departments where it reduces the
time spent on answering relevant cases by up to 60%.
The Prime Legal AI legal data model is being constantly
improved by growing content. Each law firm can use the
platform exclusively with their legal content. 

Watson saves up to 60% time

for lawyers and in-house lawyers
and helps them to find the
right answer on repetitive legal
questions

16
© IBM Corporation 2019

Authors

Dr. Stefan Mück

IBM Distinguished Engineer and Executive Partner, CTO
LinkedIn: www.linkedin.com/in/stefan-mueck

Varghese Abraham
Associate Partner ,Cognitive Process Automation Offering
Leader, Cognitive Process Automation Offering Leader
LinkedIn: www.linkedin.com/in/vargheseabraham1

Markus Brandes
Executive Partner, Head of Digital & Cognitive Solutions
LinkedIn: www.linkedin.com/in/markus-brandes