Artificial Intelligence and Machine Learning Implementation Solution

This document presents a solution for implementing Artificial Intelligence and Machine Learning in the insurance industry to improve customer experience and reduce costs.

Artificial Intelligence and Machine Learning 1
ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING IMPLEMENTATION
SOLUTION
Student
Course
Tutor
Institutional Affiliations
State
Date

Artificial Intelligence and Machine Learning 2
Introduction
Today’s customers are digitally savvy with multitudinous insurance options at hand.
Gone are the days when the local insurance agent sell insurance policies on basis of personal
relationship. In the current highly competitive landscape, insurance agencies has to cope with the
continuously changing consumer needs. The traditional strategies for fraud detections, claim
management as well as customer experience have become obsolete (Kirlidog and Asuk, 2012,
pp.989-994). The current customers, however, needs access, speed and flexibility and the
traditional strategies does not meet these needs (Honka, 2014, pp.847-884). As such,
organizations which are able to implement new technology to reduce costs as they improve
customer experience will be guaranteed a considerable market share and loyalty.
In the face of the disruptive technology in the insurance industry, Bingle agency, an
organization which operate in one of the most critical industries still lags in implementing the
current technology. In this rationale we seek present a solution to the technology proposed in the
previous article. Artificial Intelligence AI and Machine Learning is becoming the only way to
survive in this competitive environment. The sector is ready for change (Hall, 2017, 243;
Agrawal, Gans, and Goldfarb, 2018. pp. 04-21). Besides, the organization has a goldmine of
data ranging from the telematics, access to third party sources to historical data. The data can be
harnessed to drive a considerable outcome that will help in increasing the organization’s market
share and profitability. Artificial Intelligence driven by data science as well as machine learning
is the key to unlocking the available data and solving the problems which are currently faced by
the organization. However, before we venture into this critical aspect that promise a considerable
outcome to the organization, it is essential that we gain some insight into the insurance customer
journey.

Artificial Intelligence and Machine Learning 3
The Customer Journey
Analyzing the needs and expectations of customers and selecting the best mix of
marketing channel appropriately particularly in motor insurance providers is a critical aspect that
should be given attention in an enterprise solution like this (Nelson, Peterson, Rariden and Sen,
2010, pp.30-41; Markic, I., Stula and Maras, 2014, pp. 1118-1123). The customer journey will
be analyzed on basis of four broad stages including customer acquisition, premium payments,
claims processing as well as customer renewal and retention.
Customer acquisitions
This stage involves customer evaluation on premium quotes and coverage. This is an
important opportunity that shows how easy it will be to do business with the Bingle agency. The
phase is very critical in reducing leakages of prospect customers (Lemon, and Verhoef, 2016,
pp.69-96). The new system should therefore be user friendly and easy to navigate (Tax,
McCutcheon and Wilkinson, 2013, pp.454-470). This has to be considered to enhance the
consumer relationship. A proactive engagement channel is essential to encourage customers to
complete their transactions.
Premium payments
This a critical phase of journey that need its own attention. Today’s customers are
digitized. To begin with automated payments sign ups, this process can be made easy through
virtual assistance. This process will enable a timely revenue stream as it reduces delinquency and
collections.
Claims and processing

Artificial Intelligence and Machine Learning 4
Customers claims needs to be addressed as fast as possible. As the initial applications
always demand that a call into a contact center is made, the system automation will be designed
to enable an efficient and effective call routing. NLU may be considered to allow consumers to
bypass the frustrating menus, direct dialogs and key pad presses with conversational languages
(Tur, Jeong, Wang, Hakkani-Tür and Heck, 2012, pp.07). This therefore calls for virtual
assistance.
Customer renewal and retention
With the completion in this industry, there is no guarantee that a customer may return for
more services, customer renewal and retention is the most critical phase and should be accorded
the attention it deserves if Bingle is to retain its customers.
The information and technology architecture
In a description of the information architecture, for Bingle to meet the demands regarding
customer acquisition, the organization’s new online platform will be made such that it can be
easily understood and navigate by its customers. We will use artificial intelligence enabled
virtual assistance to help in handling questions, to provide quotes and even steer consumers
towards the resources. In many occasions, customers will want to call in order to verify quotes
they have been given online, we will enable this by utilizing the proactive engagement such as
automated text or emails, and even automated voice to help in confirmation, this can also help in
deflecting calls. On the off chance that the call is received, we will use a natural language
understanding NLU to assist in containing the calls thus boosting satisfaction (Khayrallah, Trott,
and Feldman, 2015, pp.94).

Artificial Intelligence and Machine Learning 5
To meet the demands concerning premium payments, the interaction voice response IVR
can make it easy by enabling a timely revenue stream (Hsieh, Kung, Lai and Chen, 2011, pp.
1000-1005). For the customers who are not on automated payments, we will use automated
reminders that will be facilitated by the customers’ channel of voice which will make it easy for
the customers to take actions from their PCs or mobile phones.
For claim processing, using the IVR designed for effective call routing, the customers’
claims will be addressed immediately they are ‘behind the login’ or send automated claims
concerning status updates. We also consider the customer renewal and retention by enabling a
renewal reminder to customers. It is recommended that we consider a customer survey for
gaging the consumers’ satisfaction with the organization and its services before this phase is
executed.
Justification
The proposed artificial intelligence virtual assistance will help in automating many
operations that would be otherwise done by human beings, this is essential as it reduce the cost
of operation. The IVR, on the other hand will ensure customer engagement which is essential to
improve the customer relationship (Korb and Nicholson, 2010, pp.3). Overall, the proposed
information and technology architecture are essential as they promise customer relationship as
they reduce cost due to automation which is the goal of this solution.
The governance methodologies that will be used in the project
Governance
This process will involve the use of customers’ data and there are chances that data
breach may occur during the process. To avoid this, we will use data governance policy to ensure

Artificial Intelligence and Machine Learning 6
maximum security; the data governance procedure will be priority so as to avoid any potential
breach (Calo, 2017, p.399). This is to ensure that the customer’s right to privacy is not violated.
Justification: protecting customers right to privacy is essential to keep the trust; on the off
chance that data breach occur, the rust may get eroded which might lead to negative impacts to
the organization.
Methods
This implementation will be done by consideration the following methods:
Provide a clear definition of the use case: this will be the first step in the
implementation. Here, we will clearly articulate the issues within the organization’s
business that has precipitated the need for AI implementation (Mitchell, Michalski and
Carbonell, 2013, pp.51). Some of the rationale for this project is to enhance the
organization’s competency in the competitive market.
Verification of the availability of data: this is to ensure that the underlying system can
effectively capture and track data that will be needed while performing analysis. In this
phase, it will be important to collect the relevant amount and volume of data that is
enough for the project.
Conducting basic data exploration: in this step, we will conduct a quick data
exploration, this is to validate the data assumptions and understanding.
Outlay methodology for building model: it is important to concentrate on hypothesis
instead of focusing on the end goal. This will validate hypothesis and improve the
execution process.

Artificial Intelligence and Machine Learning 7
Outlay model validation strategies: this method involves the definition of the
performance measures. It will help in evaluation, comparison as well as the analysis of
the results with the help of various algorithms which will help in arriving at a more
refined model. Here, data will have to be subdivided into two datasets including training
datasets which will enable training of algorithms and test datasets, a set on basis of which
evaluation will be made.
Automation and production rollout: after building and evaluating the product model, a
production rollout will be the next step. We will start with a limited rollout of a few days
as we seek for customer feedback concerning the model behavior.
Updating the model: after deployment of the model for use, the model will be monitored
on a continuous basis as required. The model should be updated frequently as it can be
out of date for various reasons such as consumer expansion and technological innovation
among others. The model will be built on basis of the historical data so as to predict the
future outcomes and also, the market is dynamic this therefore callas for the need of
update.
Justification
This methodology is the most recommended for the organization because of the
following reasons:
Having a clearly defined use case will enable us to concentrate on the specific problem
that is experienced by the agency.
Data verification is essential as the quality of data is essential for a successful outcome as
it will enable us to select the recommended data type and volume for the project.

Artificial Intelligence and Machine Learning 8
Data exploration is promising as it will help us to understand the essential variables and
features as well as the data categorization that should be modelled for use in the model.
Outlaying model building strategy will help in validating hypotheses essential for
improving the implementation.
Defining a model validation strategy will help in gaging the whether things are moving in
the right direction or not, this is essential as it will ensure effectiveness and efficiency of
the model.
Automation and production rollout will help in experimenting the model and find out if it
will solve the needs for which it is designed.
Continued update is important as due to the fact that the dynamic market keeps moving,
there are chances that the performance of the model can deteriorate (Guelman, 2012,
pp.3659-3667). It is therefore important to be mindful of the processes that will ensure
that the model is updated every time.
Work breakdown and work package decomposition

Artificial Intelligence and Machine Learning 9
Figure 1: Work breakdown and work package decomposition
The critical success factors for the project
There are five main factors that the organization must leverage to ensure the success of the
project, these include the following:
Defining the AI goal: this is one of the critical factors that should be considered during
the project. It is important to know the exact business problem that project is intended
for.
Supporting the goal with data: this is one of the great parts of AI projects that will
contribute to the project’s success.
Experts: the experts are critical to AI implementation. They can help in identification of
the good data and tell how to utilize the data in an effective way.

Artificial Intelligence and Machine Learning 10
A measure of the AI outcome: this is an important factor to the success of the project.
The AI impacts are essential aspects in understanding the success of any initiative.
Iteration: this is another considerable factor that may contribute to the success of the
project as through iteration, we will better understand the algorithms we work with, it
will also enables solving bigger problems with machine learning, it also enables
anticipation of a realistic timeline for the project among others.
The impacts of implementation of the Artificial Intelligence and Machine Learning
This implementation will demand good deal of supporting technology, as well as the
technology landscape of the insurer. The current state of the technology landscape is one of the
attributes that the organization will need to evaluate. Along with this, the state of digitization and
the level of integration between various system as well as the availability of data from various
sources are some of the critical factors that Bingle agency will evaluate before adopting the AI
(Celi, Mark, Stone and Montgomery, 2013, pp.1157; Scherer, 2015, p.353). However, as the
organization’s system is out of date, Bingle corporate must just plan for a new system and setting
up such structure as well as enabling integration with Artificial Intelligence and Machine
Learning will require both cost and time consideration.
Conclusion
In summary, this document has presented a solution to the issues experienced by Bingle
agency and we can conclude that there is a considerable potential in the AI and ML ranging from
reducing the operation costs to enhancing customer experience, customer satisfaction as well as
retention. As such, implementing AI will yield profound impacts in the organization.

Artificial Intelligence and Machine Learning 11
Reference list
Agrawal, A., Gans, J. and Goldfarb, A., 2018. Prediction machines: the simple economics of
artificial intelligence. pp. 104-321. Harvard Business Press.
Calo, R., 2017. Artificial Intelligence policy: a primer and roadmap. UCDL Rev., 51, p.399.
Celi, L.A., Mark, R.G., Stone, D.J. and Montgomery, R.A., 2013. “Big data” in the intensive
care unit. Closing the data loop. American journal of respiratory and critical care
medicine, 187(11), p.1157.
Guelman, L., 2012. Gradient boosting trees for auto insurance loss cost modeling and
prediction. Expert Systems with Applications, 39(3), pp.3659-3667.

Artificial Intelligence and Machine Learning 12
Hall, S.N., 2017. How artificial intelligence is changing the insurance industry. National
Association for Insurance Policy and Research-CIPR Newsletter, 22.,pp.243.
Honka, E., 2014. Quantifying search and switching costs in the US auto insurance industry. The
RAND Journal of Economics, 45(4), pp.847-884.
Khayrallah, H., Trott, S. and Feldman, J., 2015. Natural language for human robot interaction.
In International Conference on Human-Robot Interaction (HRI).
Kirlidog, M. and Asuk, C., 2012. A fraud detection approach with data mining in health
insurance. Procedia-Social and Behavioral Sciences, 62, pp.989-994.
Korb, K.B. and Nicholson, A.E., 2010. Bayesian artificial intelligence. CRC press, pp.3.
Lemon, K.N. and Verhoef, P.C., 2016. Understanding customer experience throughout the
customer journey. Journal of marketing, 80(6), pp.69-96.
Markic, I., Stula, M. and Maras, J., 2014, May. Intelligent Multi Agent Systems for decision
support in insurance industry. In 2014 37th International Convention on Information and
Communication Technology, Electronics and Microelectronics (MIPRO) (pp. 1118-1123). IEEE.
Mitchell, R.S., Michalski, J.G. and Carbonell, T.M., 2013. An Artificial Intelligence Approach.
Springer, Berlin, pp.51.
Nelson, M.L., Peterson, J., Rariden, R.L. and Sen, R., 2010. Transitioning to a business rule
management service model: Case studies from the property and casualty insurance
industry. Information & management, 47(1), pp.30-41.

Artificial Intelligence and Machine Learning 13
Scherer, M.U., 2015. Regulating artificial intelligence systems: Risks, challenges, competencies,
and strategies. Harv. JL & Tech., 29, p.353.
Tax, S.S., McCutcheon, D. and Wilkinson, I.F., 2013. The service delivery network (SDN) a
customer-centric perspective of the customer journey. Journal of Service Research, 16(4),
pp.454-470.
Tur, G., Jeong, M., Wang, Y.Y., Hakkani-Tür, D. and Heck, L., 2012. Exploiting the semantic
web for unsupervised natural language semantic parsing. In Thirteenth Annual Conference of the
International Speech Communication Association., pp.7.