Technology Driven Strategies: Artificial Intelligence in Business

This report examines the application of Artificial Intelligence (AI) in enterprise management, focusing on its strategic implications and practical implementations. It begins with a definition of AI, highlighting its processes of learning, self-correction, and reasoning, and its value in enhancing ...Read More

Artificial Intelligence

Introduction
Artificial intelligence is defined as replication of the processes of human intelligence by
machines, mainly the computer systems. These processes involve learning (the acquiring of rules
and information for utilizing the information), self-correction and reasoning (applying the rules
in order to arrive at approximate or certain conclusions). Artificial Intelligence can forecast when
a sensor of a key in a machine requires to be replaced to evade a shutdown on the manufacturing
line or can be utilized in braking systems that require urgency to stop robots from destroying
their own components significantly. It can predict when units will sell out in a business, provide
an overview and also respond to supply chains patterns, and even recognize factors that are risky
in investments centered on the loan repayment behavior of business and usage of the credit.
Artificial Intelligence (AI) stimulates value within businesses, particularly in customers’
experience, reducing cost and in initiatives for revenue generation. Chief Information Officers
(CIOs) who are successful are aware that the technologies involved in Artificial Intelligence
involve more than just being tactical and technical. They also understand that applying AI
successfully can open up new opportunities and assist in attaining goals of the businesses.
Majority of AI definitions emphasize on automation which in return fails to give a clear picture
of the opportunities presented to business leaders and IT.
Past, present, and future of Artificial Intelligence
Although artificial intelligence concepts have been there for centuries, its possibility was
explored in the 1950s. Expert systems were then brought about which imitated processes involved
in decision making of a human expert. However, it was in the 2000s that numerous landmark
goals have been achieved and artificial intelligence has grown despite lack of funds from the
government and attention of the public.
According to a report by Gartner, over the past 4 years, enterprise use of AI grew by
270%. A numerous number of companies are implementing artificial intelligence at a very high
rate. The 2019 CIO Survey of the Stamford firm which comprises of more than 3,000 executives
in 89 countries discovered that the implementation of AI grew 270% in the past 4 years and 37 %
in last year1. According to Chris Howard, presently we are still far from general artificial
intelligence which can entirely take over tasks that are complex. However, we have immersed in
1 Geetha, International Conference on Automation, Computational and Technology
Management, 370.

the realm of work that is AI-augmented and decision science which is referred to as augmented
intelligence2.
It is anticipated that, if CIOs consider the following things, they will contribute in speedy
evolution of AI techniques and tools and the way they will work out in their organization. The
following evolutions are to be anticipated in the future;
 Infrastructure decisions will be driven by AI
Due to the increase in the use of artificial intelligence, it's anticipated that by 2030, AI will be
among the main workloads that determine the decisions to be made regarding infrastructure.
Increasing adoption of AI needs certain infrastructure resources which can develop and evolve
along with technology. The models applied in artificial intelligence will require to be refined
periodically by the IT enterprise team to guarantee high rates of success.
 Use of collaboration to manage AI techniques complexity
It's predicted that 40% of the I&O teams will utilize the automation that is AI-augmented in
large businesses leading to high productivity in IT. This will be made possible by solving the
problem of leveraging techniques used in AI, which is the complexity of analytics and data
through collaboration. This is evident based on the fact that traditional use cases of AI that do not
include expectations of the customer are effective due to the tight collaboration between IT
functions and the business, therefore, obtaining the assistance of the engineering team internally
is a compulsory fact.
 Use of simple learning techniques
Over 75% of enterprises by 2022, will apply DNNs for use cases that possibly will be
covered through the use of ML techniques that are classical. Research states that classical
techniques of machine learning are awfully underrated3. If one carefully examines the increasing
use of AI, it’s evident that a number of organizations are striving to implement learning
techniques that are deep without knowledge of their application to their current initiatives. Due
to that, we can clearly conclude that simplicity is key. Leaders involved in IT should consider
learning of the various techniques available in order to appropriately deal with the business
problems.
2 Chris and Daryl, Gartner Research Note 6.
3 Sau et al., eds., Gartner G 343423.

 Application of Serverless computing
Serverless computing and containers will enable models of ML to function as independent
functions and in return operate with a friendly cost accompanied by low overhead.
A case study of how an organization is successfully using AI to gain a strategic
advantage.
A good case study of an organization that has successfully applied AI and succeeded in it
is Ocado, a British online supermarket which is the world’s largest online retailer for grocery4.
The company is using a deep neural network to help in improving the detection of fraud on its
platform.
The system is improving the rate of detection of wrongdoings by 15 times through
automatic detection. The online supermarket’s technology division has developed its own
learning based algorithm with Google Cloud and TensorFlow technology to assist in fraud
detection among petabytes of data that are sheer which it stores in the Cloud of Google. The
system is a section of the OSP (Ocado Smart Platform) and tries to distinguish between
customers who input wrong personal details by accident or those who are using debit cards that
are expired and those whose intent is malicious.
According to Ocado management team, the danger involved with fraud is that, if left
unchecked it could lead to moving of fraudulent information to other divisions and systems in
the business and finally affecting the services offered to customers. As a result, Ocado has
developed a mechanism that predicts and recognizes these cases among numerous of other
events that are normal through the use of the collected data from the processed orders as well as
fraud cases.
In order to achieve the intended function of the fraud detection system, the engineers at
Ocado have developed a deep neural network TensorFlow and have managed to upload the
whole system of fraud detection to the cloud. The system was built because it was accurate and
its deployment into production is easy. Two of the main reasons as to why the system was
considered viable was because of its adaptability and speed.
Machines are considered to be more capable and quick when it comes to pattern
detections when compared to humans. In addition to that, fraudsters are known to change their
4 Dominik and Reinhard, Digital Policy, Regulation and Governance.

tricks and machines are capable of learning the new patterns quickly than humans. Ocado’s
efforts in including more features in data training have led to the creation of a model that is more
reliable that has already accomplished substantial success. This has led to improvement in
Ocado’s accuracy in fraud detection by a factor of 15.
Business Case
The National Australia Bank is one of the publicly listed companies in Australia.
Artificial intelligence is readily applicable in the organization since it can be used to offer
services to both the customers and the management itself. The bank incorporated in its system
some artificial intelligence features, the latest one being a proof of concept automatic teller
machine. The system is designed by putting together artificial intelligence and cloud computing
technologies. It enables a customer to withdraw cash from an ATM using just facial recognition
and a pin thus doing away with the need to use physical cards or devices to access their cash.
With artificial intelligence, the bank can develop other systems to aid in their customer
services. Artificial intelligence can be used to develop a speech recognition system that will be
able to communicate with the customer. Through learning and experience, it will have to
understand their natural language and also interpret their question and provide solutions. One of
the features of the system will be to help a customer in opening an account, that is registering a
customer to the bank. The system will also be able to respond to customers in case they require
their account statements, depositing and withdrawing as well. The AI machine will communicate
directly with the customer and gather their information and execute the intended functions.
In the case of registration, the system will ask the customer to provide his details, through
speech and then capture them and store in the database thus there must be a link to the database.
It should be able to detect an error in the customer’s details and correct them with the aid of the
customer. The system can also provide guidelines and instruction to customers such as the
procedure required to do a certain task and also teach the customers on some important features
of the bank. It can also be trained to advertise the bank’s services to the customers once they step
into the bank through speech.
The AI machine will also have queries feature where it will be able to take in customers
question and give them immediate feedback. Customers will give their inquiries or complains
and the system can give an answer or solution where appropriate. The system can also aid

customers in booking appointments with the bank managers, through its speech feature, it can
communicate to them whether the manager is available or not and also give a hint on his free
time. The management can also use the system to their advantage by using it to pass information
to the employees such as scheduled meetings.
In order to achieve this, the bank must lay down a strategy, first is to train the system on
speech recognition it a way that it will understand the natural language spoken by humans.
Secondly is to train the system to communicate with the customers in an understandable
language, the best being a natural language. The system must be trained to store data or
information given to it by the users. The users, in this case, will be the customers, employees and
the administrators. The system will then have to be fed with the banking procedures to aid in
instructing the users. It will also be fed with possible and commonly asked questions and their
solutions.
The system will benefit all the users. The bank will benefit through its reduced manpower
since the system will be able to perform tasks that could have been performed by persons. It will
reduce the cost of hiring persons to do the work and also improve customer services since the
machine will not tire like humans. It will also reduce errors since it will be programmed only to
enter and store correct data either from the administrator or from the customers. To the
customers, it will be easy to acquire services since queueing will be reduced. The system will
also save a lot of time during the processing of customers services since everything will be
automated.
The system can be implemented through the use of as many devices as possible which
will be placed at different strategic places in the bank, or even offices making it easier to serve
multiple customers at a go. Implementing the technology will require the banks ICT team to
develop the system first or to pay outside developers to do so for them. The bank will then have
to purchase user devices such as microphones and speakers depending on the number of desired
access points. The microphones will be input devices and the speakers will be output devices.
The cost of purchasing the system and the input and output devices is sensible compared to the
services that the system will deliver and also the money used to hire persons to do the job. The
system does not go against the culture of the organization and its customers since it only
improves the quality of services and also fastens feedback from all users. With the use of

artificial intelligence, the organization will benefit financially and also gain customer satisfaction
which will greatly boost their business5.
Input and output devices
Device Description Quantity Cost Total cost
Speaker Multimedia
speaker
5 $300 $1500
Microphone Generic
condenser audio
microphone
5 $250 #1250
Desktop  HP
 8gb
RAM
 1 TB
Hard
disk
storage
 CPU
core i7
5 $270 $1350
Total $4100
System
Task Cost
Development $1000
Testing $600
Maintenance $400
Total $2000
Conclusion
Clearly, artificial intelligence is taking over business enterprises. Organizations are
rapidly adopting the technology; each year the rate is higher than the previous. The technology
has enabled enterprises to increase productivity, made management easier through automation,
helped in predicting future trends in the market which enables organizations and other people to
stay alert. Transactions are also easily carried out in banks which has made it possible for firms
such as Baton to accelerate its payments, therefore, generating more income.
In years to come, almost all important roles in the firms will be handled by artificial
intelligence technology. This will improve the effectiveness of the tasks carried out by firms.
5 David, U.S. Patent 7,634,662.

Firms with artificial intelligence have an upper hand in the market which is why the rate at which
AI is being adopted will continue increasing. With continuous learning and experience, almost
all roles will be taken over by robots and other technologies involved in AI. What needs to be
done right now is more research and learning by CIOs in order for them to stay informed of the
techniques available in the market so that when a situation arises in the organization, they will be
able to make informed decisions. Unknowingly, CIOs implement complex algorithms while
there are simple ones out there, the only difference is that they are not informed.
References
1. Sethu, Sagee Geetha. "The Inevitability of an International Regulatory Framework for
Artificial Intelligence." In 2019 International Conference on Automation, Computational
and Technology Management (ICACTM), pp. 367-372. IEEE, 2019.

2. Howard, Chris, and Daryl C. Plummer. "The Nexus of Forces Works Its Way into the
Enterprise." Gartner Research Note 6 (2013).
3. Andrews, W., M. Sau, C. Dekate, A. Mullen, K. F. Brant, M. Revang, and D. C.
Plummer. "Predicts 2018: Artificial Intelligence." Gartner (13 November). DOI=
http://dx. doi. org G 343423 (2017).
4. Weber, Felix Dominik, and Reinhard Schütte. "State-of-the-art and adoption of artificial
intelligence in retailing." Digital Policy, Regulation and Governance (2019).
5. Monroe, David A. "Method for incorporating facial recognition technology in a
multimedia surveillance system." U.S. Patent 7,634,662, issued December 15, 2009.