INTRODUCING INTELLIGENT AGENTS TO OUTSOURCING

Hemal Kothari, Bernadette Sharp, Luke Ho, Anthony Atkins

Staffordshire University, Faculty of Computing, Engineering and Technology

Octagon, Beaconside

Stafford, ST16 9DG

Keywords: Intelligent Agents, Outsourcing Decision-Support, Conceptual Agent model.

Abstract: In the last few years, agent technology has significantly emerged as a new paradigm for software developers

to solve complex problems. This paper extends the use of multi-agent systems into a new domain of

outsourcing. It highlights the various issues associated with outsourcing decision making i.e. the complexity

and the risks involved in outsourcing. The paper outlines the HABIO framework which proposes a tri-

perspective approach focusing on the organisational, information and business perspective to facilitate the

outsourcing decision-making and formulating an effective outsourcing strategy. The main focus of this

paper is to describe how agents can assist the experts in their decision to support outsourcing. A call-centre

scenario illustrating a 3-layered agent architecture is proposed which aims to capture the strategic, tactical,

and communicational layers of outsourcing and supports the experts in their outsourcing decision-making.

1 INTRODUCTION

Outsourcing is one of the most prevalent trends in

the IT industry today. It accounts for nearly over

35% (i.e. a third) of the IT and Consulting Services

market worldwide (Andersen, 2004). In the last

decade, companies are focusing on what is deemed

as the best/smart/right-sourcing of various business

functions rather than the complete Outsourcing of IT

operations which was initiated in the late 1980s by

Kodak Eastman (Loh and Venkatraman, 1992).

There are various definitions of “outsourcing”

such as “procuring of services or products from an

outside supplier or manufacturer in order to cut

cost”. In business context, it can be defined as the

transfer of an internal non-core function of a

business organisation to an external vendor.

According to Laudon et al. (2004) outsourcing can

be defined as “the practice of contracting computer

centre operations, telecommunications network, or

application development to external vendors”. Elliot

and Torkko (1996) define outsourcing as “a

conscious business decision to move internal work

to external suppliers”. Initially motivated by

necessity and later a cost-cutting move (as shown in

Table 1), outsourcing has now evolved to become a

management strategy in various industries mainly in

reducing operation costs.

Table 1: Evolution of Outsourcing.

Time Issue Forms of

Outsourcing

1960s Cost of

mainframes

Time-sharing

(e.g. payroll

processing)

1970s Cost of software

development

Contract

programming

1980s Plummeting cost

of computing

capability

Reverting to

Insourcing /

In-house retention

1989s Kodak Eastman

Mega-deal

Reactionary

trigger

(sparked

bandwagon effect)

1990s Strategic concerns Onsite facilities

management and

Selective

Outsourcing

2000s Transformational

concerns

Business Process

Outsourcing and

Offshore

Outsourcing

Although there are several advantages associated

with outsourcing (Davidson, 2004; Kruse and Berry,

2004; Namasivayam, 2004; Sloper, 2004), like any

other business operation, there is a degree of risk

involved, such as the loss of strategic direction and

316

Kothari H., Sharp B., Ho L. and Atkins A. (2006).

INTRODUCING INTELLIGENT AGENTS TO OUTSOURCING.

In Proceedings of the Eighth International Conference on Enterprise Information Systems - AIDSS, pages 316-322

DOI: 10.5220/0002447703160322

 SciTePress

loss of organisational competencies (Aubert et al.,

1998; Earl, 1996). In its 2004 report, the research

group Gartner indicated that as many as 80% of

outsourcing deals are unsuccessful and that

European businesses wasted $7bn on poorly

managed contracts (Glick, 2004). This highlights

two main issues - the first being the potential

complexity of outsourcing decision-making and the

second being the financial risks involved. One risk

which companies are increasingly concerned about

is the lack of flexibility in long-term contracts. This

was highlighted by the collapse of the Royal Bank of

Scotland (RBoS) 10 year $1.3bn mega-deal with

IBM, which was indicated to have failed due to its

inflexibility to accommodate the business changes

following the merge of RBoS with Halifax.

Outsourcing can be seen as a potentially complex

decision-making process, which ideally takes into

account various degrees of information from a wide

range of domains such as technological, economical,

geographical and political factors. Decision-makers

not only need to critically analyse all aspects of the

business, but also need to interact with multiple

organisational layers in order to understand the

implications and risk associated with outsourcing.

Outsourcing is really a huge domain covering

various aspects of a business for example human

resource management, administrations, finance,

manufacturing, insurance, credit checks, food

catering, customer service, etc making it practically

impossible to cover all these areas. The focus of this

paper is to describe how agents can assist call centre

outsourcing and integrate the various kinds of

knowledge associated with the decision making in

these operations. Due to the myriad of potential

risks, coupled with the need to incorporate multiple

facets of information, it is anticipated that the use of

strategic framework techniques, enhanced by agent

technology, will be an increasingly feasible

candidate for decision support in the outsourcing

decision-making process. The role of the agents

would focus on strategic, operational and tactical

issues within an organisation taking into account the

various facets of outsourcing. These agents will also

be managing and monitoring the service level

agreements (SLA) and risks at different stages of

outsourcing.

2 HABIO FRAMEWORK

DEVELOPMENT

Although there have been projects and initiatives

associated with Outsourcing, they tend to focus only

on specific aspects such as partnerships (Grover et

al., 1996), performance (Loh and Venkatraman,

1995) and whether to insource or outsource (Meyer,

1994). There is still a lack of an integrated method

which considers the business organisation as a whole

in Outsourcing decision-making. An ideal method

would gather, model and represent the information

in such a way which facilitates both Outsourcing

decision-making and the formulation of an effective

Outsourcing strategy. In order to facilitate this, the

proposed agents will take a Holistic Approach

{Business, Information, Organisational} (HABIO)

framework (Ho and Atkins, 2005) which addresses

issues within multiple domains, such as financial

costing, performance benchmarking and degree of

union pressure.

Figure 1 illustrates a conceptual model of the

HABIO framework. Its development was based on

the Information Systems Strategy Triangle (ISST)

which is both a well-documented concept and a

well-known convention (Frenzel and Frenzel, 2004;

Robson, 1997). The framework proposes a tri-

perspective approach to Outsourcing decision-

making which addresses issues within multiple

domains, as follows:

 The Organisational Perspective involves the

consideration of organisational issues and

political influences, such as the degree of

pressure from trade unions and internal policies

which constrain the scope of decision-making.

 The Information Perspective involves the

consideration of technical issues, such as quality

of service (i.e. fitness for purpose), performance

metrics for service benchmarking and

availability of required external expertise.

 The Business Perspective involves the

consideration of business issues, such as the

financial costing of tangible (e.g. product

delivery) and intangible (e.g. research and

development) activities within the organisation.

Akin to the ISST concept, the HABIO framework

advocates that the corners of the triangle are

interlocking, thus a change in one perspective will

require re-evaluation of the other two perspectives,

in order for corresponding adjustments to be made.

This reflects the complex nature of outsourcing

decision-making which involves multiple factors

that can influence one another, both in an intra-

perspective and inter-perspective manner, as

illustrated in Figure 2.

In order to address the complexity from multi-

influence and interrelations, it is proposed that a

fuzzy logic approach involving multi-agent systems

be utilised in the decision-making process.

INTRODUCING INTELLIGENT AGENTS TO OUTSOURCING

317

Figure 1: Holistic Approach {Business, Information, Organisational} (HABIO) framework.

Figure 2: Example scenario of inter-perspective and intra-perspective influences of factors.

3 INTELLIGENT AGENTS

The focus of this research is the use of multi-agent

system to assist in the analysis of information

collated from the various domains associated with

outsourcing and support outsourcing decision-

making.

3.1 Agent Definition

There is no universally accepted definition of the

term ‘agent’ as it depends on how this agent is used

in an application. For some applications, agents can

be defined as computer systems that have the ability

to learn from experiences whereas for other

applications, learning is not only unimportant but

also undesirable (Weiss, 1999). There is indeed an

ongoing debate and controversy to define ‘what is’

and ‘what is not’ an agent as currently every system

that performs a function is labelled as an ‘agent’.

Maes (1995) defines agent as “computational

systems that inhabit some complex dynamic

environment, sense and act autonomously in this

environment, and by doing so realize a set of goals

or tasks for which they are designed”. Franklin and

Graesser (1996) defines it as “a system situated

within and a part of an environment that senses that

ICEIS 2006 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

318

environment and acts on it, over time, in pursuit of

its own agenda and so as to affect what it senses in

the future.” In the simplest term, agent can be

defined as a software entity that is capable of

carrying out flexible independent activities in an

intelligent manner in order to accomplish tasks to

meet its design objectives, without direct

intervention and guidance of humans or other

applications or agents.

3.2 3-Layered Agent Architecture

Though cost is an important viable business strategy

to outsource within a call centre, however it is not

just about cutting costs but improving the services,

focusing on the core components and improving the

ability to cope with internal and external demands.

The main issues within a call centre include

forecasting and scheduling (i.e. predicting the day to

day workload and calculating the staffing levels

accordingly), and managing the service levels

(achieving the targets, for example, answering a

certain percentage of call within a certain number of

seconds). In order to successfully achieve the set

targets, the agents need to extract the right

information at the right time from the right experts

and present to the right people. In order to

effectively assist project managers in outsourcing

decision-making, agents need to reflect the various

organisational layers (typically as strategic, tactical

and operational). Figure 3 proposes a three-layered

agent architecture consisting of the strategic, tactical

and communicational layers. The strategic layer

captures the specific domain knowledge of the

project managers by means of ontological

representations and identifies the role of each agent.

As different agents have different goals, the strategic

layer therefore captures the various facets of

outsourcing, in particular financial costing (tangible

and intangible costs), risk management and

benchmarking. The tactical layer is responsible for

defining the list of actions for each agent to carry out

a particular activity while the communicational layer

is responsible for defining messages which would

enable exchange of meaningful information between

the agents. It is noted that communication cannot be

reduced to just the sending and receiving of

messages, and thus the communicational layer is

also responsible for proper interpretation of the

message, i.e. validation and corresponding

corrective measures if required.

Figure 3: 3-Layered agent architecture.

4 BENEFITS OF USING AGENTS

Multi-agent systems do not wait for instructions and

are capable of reacting to any change in situation

without being prompted. Additionally, their agents'

domain-specific knowledge is refreshed from

ontology without shutting down the system. Thus,

agents would be providing their service to the

experts 24x7 and assisting them in their decision.

Examples of how these agents will be able to assist

the experts to achieve the targets are as follows:

 Cost analysis – This forms a large part of

making an outsourcing decision and needs to be

done very thoroughly and carefully. These

agents would predict all the various types of

costs, direct and indirect, and offer a true

comparison to that of the suppliers’ costs and

therefore estimate the cost advantage that the

outsourcing could provide. Agent would also be

calculating the cost per call and cost per seat

within a call centre which would further help in

the outsourcing decision making.

 Monitoring service levels – Within a call centre,

it is well known that acquiring a new customer

is more expensive than maintaining on existing

ones. Therefore, maintaining customer

satisfaction by achieving good service levels has

a significant impact on the organisation’s

revenues. The agents will be assisting the

managers in monitoring the performance and

service quality within the organisation for

example, some of the performance

measurements include number of calls

answered, average talk time, abandonment rate

etc.

 Forecasting – Forecasting within a call centre

can be done daily, weekly, monthly, and even

yearly. Long term forecasts can be used for

budgeting, staff planning, operational changes,

training, and scheduling vacations whereas short

term forecast such as daily, weekly or every

fortnight are needed for call centre workforce

staffing and scheduling and to predict the call

volumes (Klungle, 1999). Call centres normally

Strate

Communicational

Tactical

INTRODUCING INTELLIGENT AGENTS TO OUTSOURCING

319

use manual spreadsheets for scheduling the

staffs; this approach is usually guesswork at

best, based upon the experience of the manager.

Under these circumstances it is uncommon for a

call centre to be caught flat-footed by demand

spikes or to have idle advisors (Robb, 2004).

Overstaffing and understaffing can have a huge

impact on the business perspective of the call

centre. Overstaffing can lead to having idle-

advisors which affects the financial side to the

business whereas understaffing can create a

tension environment as advisors are not able to

cope up with the amount of calls. This leads to

dissatisfied employees as well as customers.

The agents, being highly dynamic and scaleable

in nature, would be able to collate and compute

the predictions automatically taking into

consider the historical data and thus scheduling

the advisors by interacting with the various

organisational policies.

 Analysing ‘what-if’ scenario – The behaviours

of the agents are programmed with rules when

put into motion, these agents would provide

insights into many different scenarios within the

outsourcing domain. As the managers change

certain agent characteristics or rules, the

corresponding impact can be seen in the model

output thus helping the experts to take control of

various situations. For example, within the call

centres, some of the what-if scenarios that these

agents may be used to provide answers are as

follows:

o ‘What’ happens when there are abnormal call

volumes?

o ‘What’ would be the total cost of

outsourcing?

o ‘What’ will be the effect on the service level

and abandonment rate ‘If’ it is decided to

cross-train some of the advisors to take

incoming calls? How much it would have an

effect in day to day business cost?

Figure 4 gives an example of the way that the

agents would able to assist the manager in

identifying the various risks and factors involved if

they plan to outsource. The diagram indicates that

the reduction of the organisation’s profits depends

on the percentage loss of customers, forecasting of

the call volume and staffing levels.

5 AGENT MAPPING TO THE

HABIO FRAMEWORK

Outsourcing is a very complex decision making

process and takes into account the various

managerial, economical and technical issues. A

wrong decision either at the decision or at

implementation level can have devastating effects on

an organisation.

In the case of a call centre, four principal factors

have been identified, from the literature, as the most

relevant factors for outsourcing decision making.

These factors include Cost (i.e. calculation of the

total operation costs, cost reduction, etc.),

Performance (i.e. breakdown of IT performance,

improved service and quality, etc), Quality (i.e.

monitoring and improving the quality of service and

productivity), and Risks (i.e. identifying and

reducing various risk factors with the business

domain) (Alner, 2001; Clark et al., 1995, Jurison,

1995; McFarlan and Nolan, 1995).

Figure 4: Example scenario of agent-based what-if analysis in identifying the risks factor.

ICEIS 2006 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

320

COST PERFORMANCE

RIS

QUALITY

Agent A2

Agent A3

AgentA1

Agent A4

1) Hardware cost

2) Software cost

3) Utilities

)

Non-IT assets

1) Average speed of

call

2) Abandonment rate

3) Total queues

1) Staff Attrition rate

2) Disaster recovery

3) Security issues

)

islation laws

1) Call quality

2) Customer and

employee satisfaction

)

Percenta

es of re

eat

A widely known shortcoming of business

applications is the separation of analytical and

actionable processes (Politan, 2003). This creates

disconnection on both business processes and the

technical implementation both of which leads on to

various complications of the decision making within

an organisation. The agents are used to overcome

this inadequacy by its ability to automatically

translate the business intelligence into a process on

which users/agents can act. Figure 5 and 6 illustrate

these factors in a high and low level conceptual

model from agent implementation point of view

representing various agents acting together in the

outsourcing domain. The high level conceptual

model, Figure 5, defines the three mains facets of

outsourcing identified as People, Process and

Technology where people involves various experts

from which tacit knowledge can be captured by the

agents; process includes various activities which the

agents would commit to in order to meet its design

objectives; and technology defines various tools,

protocols and techniques the agents would be using

in order to interact effectively and efficiently with

various users, hardware and software applications in

order to get relevant information about the domain.

Figure 5: The 3 facets of outsourcing.

Figure 6 provides an operational view of the

agents as they are identifying the current status of

the call centre. As shown in the diagram, Agent A1

is responsible to calculate the organisation’s

operating cost (tangible, intangible and downtime

costs), Agent A2 is responsible for measuring the

current performance against the targets (set by the

experts), Agent A3 monitors the quality of service

provided to the customers, and manages the service

levels, and finally Agent A4 identifies the various

risks, reports it to the relevant experts and a strategic

approach.

Figure 6: Operational view of call centre agents.

6 CONCLUSION

This paper extended the use of multi-agent systems

into new domain of outsourcing for decision-making

purposes. In traditional approaches, procedures were

manually defined and thus required significant time

and human effort. The use of multi-agent systems

alleviates this by providing call forecasting, staffing

levels, monitoring and managing their performance.

This is achieved through the 3-layered approach

capturing the strategic, tactical, and

communicational layers of outsourcing.

The paper outlined the HABIO framework, which

proposes a holistic approach to outsourcing

decision-making that emphasises the importance of

analysing the organisation as a whole rather than

isolated analysis by parts. High and low level

conceptual models of the agent implementation were

discussed, in the context of a call centre. The various

benefits which such agent implementation could

provide to call centre operations were also discussed

in the paper.

Current research is being carried out to develop

an agent-based ontological model to emulate the

outsourcing decision-making procedures. This is

being done in collaboration with a large retail

company which operates several call centres, one of

which employs over 1500 call centre advisors.

Process Technolo

People

Strate

Tactical

Communication

AGENT

INTRODUCING INTELLIGENT AGENTS TO OUTSOURCING

321

REFERENCES

Alner, M. (2001). The Effects of Outsourcing on

Information Security. Information Systems Security,

10(2), 35–43.

Andersen, D. (2004). Proactive Planning: A Prerequisite

for Successful Sourcing. November 2004. Retrieved

January 10, 2005, from:

http://www.metagroup.com/us/displayArticle.do?oid=

49284

Aubert, B., Patry, M., and Rivard, S. (1998). Assessing the

Risk of IT Outsourcing, In: Proceedings of the 31

Hawaii International Conference on Systems Sciences

(HICSS’98), Kohala Coast, Hawaii.

Clark, T., Zmud, R., and McCray, G. (1995). The

Outsourcing of Information Services: Transforming

the Nature of Business in the Information Industry.

Journal of Information Technology, 10(4).

Davison. D., 2004. Top 10 risks of offshore outsourcing,

February 2004. Retrieved June 18, 2004, from

http://searchcio.techtarget.com/originalContent/0,289142,s

id19_gci950602,00.html.

Earl, M. (1996). The Risks of Outsourcing IT. Sloan

Management Review, 37(3).

Elliot, T. and Torkko, D. (1996). World Class Outsourcing

Strategies. Telecommunications, 30, August.

Franklin, S. and Graesser, A. (1996). Is it an Agent, or just

a Program?: A Taxonomy for Autonomous Agents. In:

Proceedings of the Third International Workshop on

Agent Theories, Architectures, and Languages (ATAL-

96). Budapest, Hungary.

Frenzel, C. and Frenzel, J. (2004). Management of

Information Technology, 4

Edition. Course

Technology, Canada.

Glick, B., 2004. Does your Outsourcing agreement

measure up? March 2004. Retrieved July 2005, from

http://www.vnunet.com/features/1153857.

Grover, V., Cheon, M.J., and Teng, T.C. (1996). The

effect of service quality and partnership on the

outsourcing of information systems functions. Journal

of Management Information Systems, 12(4).

Ho, L. and Atkins, A. (2005). Strategic Frameworks for

Outsourcing Decision-Making: The Need for a

Holistic Approach. In: H. Kehal (ed.), Outsourcing

and Offshoring in the 21st Century: A social-technical

perspective, In Press.

Jurison, J. (1995). The Role of Risk and Return in

Information Technology Outsourcing Decisions.

Journal of Information Technology, 10(4).

Klungle, R. (1999). Simulation of a Claims Call Centre: A

Success and a Failure. In: Proceedings of the 1999

Winter Simulation Conference, Phoenix, Arizona,

December.

Kruse, G. and Berry, C. (2004). Outsourcing: The “How-

To” Guide, IEE Manufacturing Engineer,

August/September, 36-39.

Laudon K., and Laudon, J. (2004). Management

Information System, 8th Edition. Pearson Education

International, New Jersey.

Loh, L. and Venkatraman, N. (1995). An Empirical Study

of Information Technology Outsourcing: Benefits,

Risks, and Performance Implications. In: Proceedings

of the Sixteenth International Conference on

Information Systems, Amsterdam, Netherlands,

December 10–13.

Loh, L. and Venkatraman, N. (1992). Diffusion of

Information Technology Outsourcing: Influence

Sources and the Kodak Effect. Information Systems

Research, 3(4).

Maes, P. (1995). Agents that reduce work and information

workload. Communication of the ACM, 37.

McFarlan, F. and Nolan, R. (1995). How to manage an IT

Outsourcing Alliance. Sloan Management Review,

36(2), 8–23.

Meyer, N. (1994). A sensible approach to Outsourcing: the

economic fundamentals. Information Systems

Management, 11(4), 23–27.

Namasivayam, S. (2004). Profiting from Business Process

Outsourcing, IT Professional, 6(1).

Politan, A. (2003). Strategic Finance. Institute of

Management Accountants, Montvale, USA, 2003.

Robb, D. (2004). Call Centre Increases Profitability Using

Workforce Optimisation Technology. September 2004.

Retrieved February 21, 2005, from

http://www.enterprisenetworksandservers.com/monthly/art

.php/933.

Robson, W., (1997). Strategic Management & Information

Systems. Pitman Publishing, Great Britain.

Sloper, A. (2004). Meeting the challenge of Outsourcing,

IEE Engineering Management, June/July, 34-37.

Weiss, G. (1999). Multi-agent Systems- A Modern

Approach to Distributed Artificial Intelligence. MIT

Press, London, England.

ICEIS 2006 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

322