INSURANCE & PAPER

An analysis towards the possibilities of digitalization and a more computerized future within the subscription process

Art. 162.1 A lot of paperwork makes your company inefficient and can eventually result in loss of market share.
Issued by T.A. Hennis & J. van Waalwijk, May 26, 2005

Chile –

University of Technology

26 May 2005

Preface
This report is a preliminary analysis about the possibilities of the Health Insurance current underwriting process of ING Chile. Computerization plans are already made and the first implementations will occur at the end of the year. During this three month project we worked at the ING head office in Santiago for the operations division. We would like to thank Francisco Javier Lopez and Jaime Ignacio Avila who helped us a lot during our project and did their upper best to understand and improve our Spanish. Also we would like to thank Professor Alexander Verbraeck for helping us out with some problems we encountered during our project. We had a wonderful time here during our stay, learned a lot about how ING really operates in Chile and got acquainted with how the underwriting process is being done.

2

Chile –

University of Technology

26 May 2005

Table of contents
Preface................................................................................. .......................2 Table of contents..................................................................................... ......3 Executive summary.......................................................................................5 Resumen ejecutivo (Castellano)......................................................................7 1.Project description.............................................................................. ........9 General outline............................................................................. .............9 Our share in this project.............................................................................9 2.Conceptualization.....................................................................................11 Process and product line...........................................................................11 Process orientated description...................................................................11 Object description....................................................................................12 Interaction-flow diagram...........................................................................12 Assumptions......................................................................................... 12 ... 3.Specification............................................................................................13 Reducing the conceptual model..................................................................13 Performance-indicators.............................................................................13 Data acquisition........................................................................ ...............14 Process times..........................................................................................14 Decisions in the subscription process..........................................................14 Routing times....................................................................... ...................14 Distributions data-input............................................................................14 4.Set-up .................................................................................................. .16 . Treatment........................................................................... ....................16 Warm up period.......................................................................................16 Replication length ...................................................................................16 Number of replications......................................................................... .....17 5.Verification..............................................................................................18 Input variable............................................................................... ...........18 System variables.....................................................................................18 Output variables......................................................................................18 6.Validation ...............................................................................................19 End of the month boom............................................................................19 7.Results simulation model...........................................................................21

Current underwriting.................................................................. .......9

Future underwriting......................................... ...............................22

8.Computerization options............................................................................22 Evaluation module.................................................................................. 22 .. Business Rules Engine..............................................................................22 Digitalize all the paperwork within the Sales division....................................22 Interaction with client...............................................................................23 Agent with iPAQ.................................................................................. 23 .. Client comes at the front desk................................................................23 Web based application ..........................................................................23 9.Options and results.......................................................................... .........25 Evaluation module.................................................................................. 25 .. Business Rule Engine......................................................................... .......25 Sales digital............................................................................................26 10.Scenario analysis....................................................................................28 Interactions.........................................................................................28 External variables.................................................................................29

3

Chile –

University of Technology

26 May 2005

Scenario 1 – The friendly sales agent.........................................................30 Scenario 2 – A bit of everything.................................................................32 Scenario 3 – ING.CL.................................................................................33 11.Summary results....................................................................................35 12.Conclusions and recommendations............................................................37

4

Chile –

University of Technology

26 May 2005

Executive summary
At the operations division of ING Chile a research is conducted towards digitalization possibilities within the underwriting process. The analysis focuses on the subscription process of the Healthcare product division, but could, because of many similarities, be useful for other product divisions as well. After a thorough understanding of the process a conceptualization and specification of all the available information about the subscription process was made. Using different techniques, the information was turned into a simplified conceptual model which was ready to be put into a simulation model. The software chosen to simulate the subscription process is called Arena and is based on discrete modeling. The program can be very helpful in discovering bottlenecks, processing times and utilizations of resources. With the available information a fairly representative model is made. This model is used to predict the consequences of certain digitalization options available for ING. Three options, which differ in degree of digitalization, have been tested. These options are the Evaluation Module (EM), the Business Rule Engine (BRE) and the digitalization of the whole sales division. The first two options do not really show promising results for the utilization of resources, but can have many advantages in for example data storage and rules management. Besides, these options are a first step towards further and more elaborate digitalization. Clearly the last one, that includes the first two options and inhibits a vast intervention within the whole process, shows a large decrease at the utilization of resources and a much smaller average subscription time. As well the effects on data storage are big, this is of great help in processes inside and outside this analysis. Two more options are part of the scenario analysis, which assumes all of the previous options implemented. Besides the digitalization (including EM and BRE) of processes, the ING-client interactions within the process can happen on different ways. These interactions can now happen digitally, with an agent using an iPAQ, type of handheld computer, connected to a server at ING, or by the client logging on to a website. These technological possibilities are examined differently in three scenarios. The configuration of these scenarios depend on three external variables, which are the use of internet, the value of personal contact between agent and client and the acceptance of digital agreements by the government. Technological possibilities, such as EM, BRE, websites and iPAQs, allow ING change their processes radically and make more efficient use of the resources. The company can make a lot more money by internally digitalizing certain processes and manage their data flows differently. Externally some processes, such as agent-client interaction, can also be changed, but are as well dependent on the described external variables. The effects of the technological possibilities are described per different future situation. In the best scenario, when things as digital signing and storage are accepted, 5

Chile –

University of Technology

26 May 2005

clients act independently of agents and the subscription is merely done over the internet (80%), the number of resources needed decreases with almost one-third and the average subscription time is halved. The client can do his/her application within half an hour, without intervention of any resource. ING can do the following to create a more efficient and apprehensible business process:  Implement the Evaluation Module and Business Rule Engine, but keep in mind that these standalone features will not bring big changes. They should be seen as a step towards a more digitally enhanced system. Conduct a better research (with more validated data) to create an improved model of the current situation. Arena is a good option to simulate the business process. Start scenario analyses for the following subjects: o Technological possibilities o Governmental changes o Marketing in the field of client’s needs. Do cost-benefit analyses for the technological possibilities.

6

Chile –

University of Technology

26 May 2005

Resumen ejecutivo (Castellano)
En la división Operaciones de ING Chile, se hizo una investigación para explorar las posibilidades de incorporar la digitalización en el proceso de suscribir. El análisis se enfoca el proceso de suscribir de Isapre, pero también podría ser útil en otros procesos o líneas de negocio, debido a las similitudes entre estos. Se realizó una conceptualización y especificación de toda la información del proceso de suscribir, después de una comprensión minuciosa del proceso. Con diferentes técnicas, la información fue cambiada en un modelo conceptual simple, para transformarlo en un modelo de simulación. El software elegido para simular el proceso de suscribir se llama “Arena” y está basado en modelamiento discreto. El programa puede ser muy útil para descubrir cuellos de botella, tiempos de procesos y utilizaciones de recursos. Se realizó un modelo bastante representativo con la información obtenida. Este modelo se usó para predecir las consecuencias de ciertas posibles opciones de digitalizaciones por ING. Fueron testadas tres opciones, las cuales son diferentes en el grado de digitalización,. Estas opciones son el Modulo de Evaluador (EM), Motor de Reglas (BRE) y la digitalización de toda el área de Ventas. Las primeras dos opciones no muestran resultados muy prometedores debido a las utilizaciones de los recursos, pero podrían tener muchas ventajas, por ejemplo, en el almacenamiento de datos o en el manejo de reglas. Además de estas opciones, están los primeros pasos a una digitalización mucho más esmerada y sofisticada. Claramente, la ultima opción, que incluye las primeras dos e inhibe una gran intervención dentro del proceso, muestra una gran disminución en la utilización de los recursos y un tiempo promedio de suscripción mucho más bajo. Además los efectos en almacenamiento de datos son grandes, lo que es muy provechoso en procesos dentro y fuera de este análisis. En el análisis de escenarios, que asume todas las opciones previas implementadas, hay dos posibilidades más. Además de la digitalización (incluyendo EM y BRE), las interacciones de los clientes con ING pueden ser de diferentes maneras. Estas interacciones pueden ser digitales, usando una iPAQ, que es un tipo de computador de bolsillo, conectado al server de ING, o a través de una página Web donde el cliente realice la suscripción. Estas posibilidades tecnológicas se han examinado en tres escenarios diferentes. La configuración de estos escenarios depende de tres variables exógenas, que son: Uso de Internet de los clientes, la aceptación de acuerdos digitales de Gobierno y el contacto personal (cliente con agente de ventas). Posibilidades tecnológicas, como el EM, BRE, Internet e iPAQs, permiten a ING hacer grandes cambios en los procesos y reducir las utilizaciones de los recursos. La compañía puede ganar más por digitalizar algunos procesos dentro de la compañía y controlar los flujos de datos de forma diferente. Pueden cambiar también algunos procesos al exterior de la compañía, como 7

Chile –

University of Technology

26 May 2005

son la interacción entre clientes y agentes, pero dependen también de las variables exógenas. Los resultados de estas posibilidades tecnológicas son explicados en las diferentes situaciones futuras. En el mejor escenario, cuando los acuerdos, los archivos digitales son aceptados, los clientes no requieren mucha ayuda de los agentes, usan mucho el Internet (80%), en ese caso el número de recursos necesarios se reduce 70% aproximadamente. El tiempo de proceso de una solicitud se reduce 50%. Un cliente puede hacer una solicitud en casi media hora, cuando la intervención de un empleado de ING no es necesaria. En resumen, ING puede hacer las siguientes acciones para crear un proceso de suscripción más eficiente y más clara: • Implementar el Modulo Evaluador y BRE, sin embargo, cabe destacar que estas implementaciones no tienen muchos cambios significativos. A pesar de esto, son el primer paso para un sistema digital que es más avanzado. Hacer un mejor análisis con datos validados de una forma más real, para crear un modelo mejor para la situación actual. “Arena” es una buena opción para eso. Empezar un análisis para las siguientes situaciones: o Posibilidades tecnológicas o Cambios en las políticas de gobierno o Marketing para determinar las preferencias de los clientes Hacer un análisis costo-beneficio para las posibilidades técnicas.

• •

8

Chile –

University of Technology

26 May 2005

Current underwriting
1. Project description
General outline A few years ago ING Chile incorporated a big competitor within the insurance business. Between 2000 and now, a lot of efforts were made to reshape processes, marketing strategies and the organization itself into one clear insurance company. Nowadays, ING is the biggest competitor in Chile regarding the insurance business with more and more clients who appeal for a new or different insurance. One of the most time-consuming processes in the insurance business is the underwriting processes. Nowadays, a lot is done manually, which causes tremendous amounts of paperwork and causes long waiting times (up to one month) for the clients and more expenses for the ING. ING decided to computerize a part of the underwriting processes. The computerization of the underwriting processes will be made in the Healthand Life insurance business. ING already chose a specific system to computerize the underwriting. The system that will be used is a Business Rules Engine (BRE). The main objectives of the BRE project are: » Reduce process times » Reduce number of resources » Improve storage of information Our share in this project Before implementing BRE it is useful to get more insight in the changes it will bring. Obviously, simulating the process of judging policy requests, which in fact will be done by a BRE-system, is not much of a use. Better is to compare the current situation with a possible future one (with BRE) to decide upon possible actions. These actions will regard the use of resources, expected bottlenecks etcetera. At this moment the BRE project is still in its initial phase and they have not agreed yet in with specific parts of the underwriting process the BRE will be implemented. Our objective in the BRE project is equivocal. First it is important to get a clear understanding of the underwriting process at this moment. It is necessary to identify important variables and obtain information about the current performance of the system. A simulation will be made of the underwriting process, which will represent the current processes. Secondly it is important to obtain more insight in the future situation. This will encapsulate the creation of some possible scenarios, in which digitalization of processes will play a significant part. ING is primarily 9

Chile –

University of Technology

26 May 2005

interested in obtaining information about possible outcomes of digitalization alternatives, which will be possible in the (near) future. Therefore, after the conceptualization and simulation of the current situation, a number of realistic and futuristic scenarios will be written. These scenarios will be simulated and judged according to the performance-indicators. The report deals with the effects of possible computerization alternatives. Therefore the problem definition can, within the set boundaries of process and product line, be described as follows: ‘What influences do future computerization alternatives have on the underwriting processes, the resources needed and the time it takes for a request to be handled?’ In the next two chapters the whole conceptualization and specification are taken into consideration. This is followed by the configuration of the simulation, together with its verification and validation. The last chapter of this part deals with the results of the simulation model. The part that follows after this deals with future underwriting, where different possibilities are being described and simulated. The results of these simulations will be judged and summed up in the conclusions and recommendations.

10

Chile –

University of Technology

26 May 2005

2. Conceptualization
In the conceptualization a conceptual model of the underwriting process will be written. The demarcation describes the processes and parts that will be left out and which part of the problem description will be focused on. Because there are a lot of different processes, resources and variables that are part of the system, some techniques were used to list all the different aspects of the entire process. After a short description of the area where the analysis will focus on a process orientated and object orientated description will be used to map the process in further detail. Process and product line Not the entire underwriting process will be part of the analysis. ING chose to focus on the subscription of new contracts, which, at the moment, is a time and resource-intensive process. Within the process of subscription, the product line that will be evaluated is that of Healthcare within the whole of Chile. Everything but the insignificant parts of the process that falls between the initial request, either by call of insurance agent or client that visits client, and the rejection or acceptation of the policy, will be handled in this report. All the elements within it, the processes and resources, will be part of the analysis of future alternatives. The whole process is divided into two parts, because of efficiency and clarity reasons, which are linked subsequently. Both parts are of the same process and should not be seen as separate processes. Process orientated description The flow of the request, from start to end, consists of numerous different processes, done by different resources. In short it can be described as follows: ‘An initial request is done, either by a client visiting an office or by an agent selling a client a policy by phone and visiting the client. After a first judgment of the policy there are several possibilities for the request, such as a more thorough medical control, a revised policy proposal or rejection. The medical control as well can have different outcomes such as a request back to the client to hand over more medical background. When a request is not rejected several documents are filled out and signed by the client. It then, after a check whether the request is not due, flows back into he sales department for digitalization. The requests that are due (enter after a certain date) are handled manually at control production. After a notification of the employer the request is signed and either approved or not. If the request is not approved, the client has to hand over the legal 11

Chile –

University of Technology

26 May 2005

documents that are provided earlier. Finally all the requests will be examined in a Production Control process. If there are errors in the documents they are revised here and the requests that were due are notified by the employer. The output of this final process are approved requests sent by mail or rejections.’ The description above is only a very minimal overview of the processes. In appendix 1 and 2 the subscription process is handled more elaborately. IDEF0 diagrams in combination with small explanatory stories are used to make a clear overview of the processes, its in- and outputs. They as well show the resources that are needed and the controls that have an influence on the processes. Object description This part regards the conceptualization of the different objects that are elements in the system. These objects have attributes, and describing these objects provides a clear input for the model. The objects can be passive or active and sometimes have actions. They are classified into different object classes. The list of objects can be found in Appendix 3. Interaction-flow diagram Using the previous analysis a better overview of all the processes, the flows and the interactions can be made. After numerous round-the-table conversations with experts, consultations, questions and feedback, a fairly realistic interaction-flow diagram is made. It describes the flows of requests and the interactions between different resources and clients and can be found in appendices 4 and 5. The diagram shows (all) the processes in the right sequence, the flows, feedback flows and decisions. It will serve, together with the obtained data (specification, next chapter), as primary input for the simulation. Assumptions Not all of the obtained information is useful when making a simulation, some aspects have to be simplified. This is necessary to make the model as clear as possible, which has a lot of advantages when adjusting it to for example future situations. Only insignificant parts and variables within the system are left out in order to create a simple and realistic representation. Besides simplifying the situation, it is sometimes necessary to do some assumptions, when things are not clear or information is lacking. This is, just as the rest of the conceptualization, an ongoing process. These assumptions and demarcations are to be found in Appendix 6.

12

Chile –

University of Technology

26 May 2005

3. Specification
The conceptual model, as described in the previous chapter, with all the described elements, objects and processes, provides a good overview of the problem situation. Nonetheless, no numbers or data have been assigned to it, leaving it an empty overview. In this chapter the formulated concepts are being simplified and data will be assigned to them. Reducing the conceptual model In order to keep the model as simple as possibly can, without loosing too much quality, insignificant parts of the conceptual model have to be reduced. Therefore objects, processes and the time-aspect can be reduced. Many assumptions and reductions are treated earlier and mentioned in appendix 6. the most important will be dealt with below. One very important object reduction regards the request. In this analysis the most important element is the request. This request is the only thing that flows through the system. Neither clients, nor agents will be part of a flow, only of processes. In reality the request can exist of different copies and forms, but in this analysis the request is the only entity and encapsulates all these possible kinds of forms, such as FUN, notification or declaration. Furthermore the time that will be analyzed regards only working days. Therefore the weekends are left out for the simulation, just as the nonworking hours. Only the lunch break has been taken into account. A day counts 10 hours, of which a normal lunch break takes 1,5 hours. Hence, a simulation week is 50 hours. Several of the processes discussed with experts have been integrated into one process. This happens with processes that are subsequent, done by the same resource and dependent of the same elements. That means that for none of the integrated processes is subject to a new element, such as an interaction with a client or another input. Performance-indicators Before we can judge the outcomes of the model that will be made or compare results later on between different scenarios, it is necessary to formulate some performance-indicators. These have been formulated in consultation with ING. The following performance-indicators will be used: Total process times: Approved requests Utilization resources: Sales Agent Sales Manager Front desk employee Controller ‘Technical Analysis’ Controller ‘Control Production’ 13

Chile –

University of Technology

26 May 2005

Data acquisition Data are obtained by consulting several experts within the company. If information was lacking, other departments of the company were consulted, such as the planification department, where data is processed and measured. Therefore, a significant part of the obtained data can be seen as expert assessments or measured data from planification. Because almost all the information flows ran through two analysts1, there is not much information about the specific origins of the data. Process times The first things to be examined are the process times and the resources doing the processes. The following resources are involved in the process:  98 Sales managers  579 Sales agents  50 Front office employees  40 Service assistants  5 Controllers (Control Production)  3 Controllers (Technical Analysis) All the process times for each individual process have been listed per resource, including its reprocess times, in appendix 7. Decisions in the subscription process This part of the conceptualization regards the different decisions within the subscription process. Within the subscription process there are many possible flows possible. Request can be rejected, incomplete, and questionable or need extra information. There are different kinds of notifications and different kinds of judgments, clients possibly need to sign extra documents and documents can contain errors and need reviewing. For all these possibilities there are several distributions, which can be found in Appendix 8. Routing times Within the subscription process there are several interactions between different parties. These interactions encompass time for sending and receiving information/ documents and visiting clients. Besides routing times, delays can be determined. These delays include for example the time it takes for a client to collect extra medical background or it takes for an employer to judge a notification. Specific information about all the routings and delays can be found in Appendix 9. Distributions data-input There are two different data-inputs in this system. One is the sales agents
1

Jaime ‘Tony’ Ignacio Avila, jaime.avila@ing.cl and Francisco ‘Pancho’ Javier Lopez, francisco.lopez@ing.cl

14

Chile –

University of Technology

26 May 2005

who call potential clients if they are interested and make an appointment to visit them. The other one is when a client enters a building of ING and requests a policy at the front desk. Clients that will be called Part of the job of a sales agent is to sell policies. They do this by calling and visiting clients. Not every call the agent makes is successful and the time simulated per called request is the time it takes for an agent to sell a policy, thus including the many unsuccessful calls. Of all the requests, 95 percent are called by the sales agent. Clients at Front desk Only a small number of clients take the initiative to visit an ING office to do an application for a policy. This number is approximately 5% of all the requests flowing into the system.

15

Chile –

University of Technology

26 May 2005

4. Set-up
After conceptualizing the situation and specifying the corresponding data a simulation model of the situation has been built. This simulation model has to be checked on whether it provides the right output, on the correctness of the model logic and on its resemblance with reality. Before that, the model characteristics itself, such as the running period and numbers of replications, have to be calculated in order to avoid fluctuating outcomes. Treatment Before the model can be verified or validated, its characteristics have to be calculated. At first the difference between an ending and a non-ending system has to be identified. An ending system is a system that has a definite start and end. This system is assumed a non-ending system, since we look at the whole subscription procedure. There are several ending systems within the whole system, such as the first subscription process at the front desk, that ends every day, but as a whole the system is identified as a non-ending one. This has certain implications for the run treatment. Three things have been taken into account with the run treatment: 1Warm up period 2Replication length 3Number of replications All the calculations and specifications can be found in Appendix 10. Warm up period The warm-up period has to do with the fact that a non-ending system does not start with zero entities in it. Therefore for a while entities have to flow through the system, until all important processes and parameters are stabilized and the flow in more or less equals the flow out, called a steadystate. From that moment on the variables in the model are calculated, so that the initial values don’t significantly influence the outcomes. The warm up period is set on 1000 hours, which is 100 (simulation) days, after which every process is stabilized, such as Control Production and the total subscription time. Replication length The replication length regards the simulation time after the warm up period. Defining this parameter depends on the variables where we are interested in. Of the variables where we are interested in enough data has to be generated to make accurate conclusions. There exists no interest in very rare occasions of requests that flow through a certain process a couple of times, has errors and needs a notification to the employer. These requests exist, but are not important in this context. Purpose is to describe the largest part of requests that flow through the system, we do not focus on special flows. 16

Chile –

University of Technology

26 May 2005

Rule for determining the replication length is to multiply the cycle with the longest average with 3. In this case the replication length is will be the three times the average time it takes for a request to be sent through the system. For average requests it takes between 110 and 120 hours to get round, or 11 or 12 working days. The replication length will be set on 42 days, which is approximately 2 months. Number of replications The calculation of the number of replications is done by means of the results of previous replications. The results of these previous replications and the calculation of the number of replication can be found in appendix 10. The number of replications is set one 1.

17

Chile –

University of Technology

26 May 2005

5. Verification
In the verification phase the variables in the model and the model structure are checked. Not the consistency with reality is checked, but the translation from the conceptualization and specification to simulation model. First the input variable will be checked, then the model-logic and finally the output variables. The output data have been calculated before the setup and originate from a simulation with the following features: 1 replication with warm-up period of 30 days and replication length 200 days. Input variable The simulation itself only has one input, which is the number of clients that request a policy each day. The average number of clients that enter the office is 346 per hour. After two replications, the results of the input were converted to number of clients per hour and are listed below. Input 59386 58472 Per hour 349.3 344.0 Average 346 346 Difference 0.96% -0.59%

Run1 Run2

Table 1 - Verification input

System variables Besides the input variables there are numerous system variables to be checked, as well numerous ways to do this. Using logic and a bit of math the system variables have been checked on consistency. Big differences between expected and calculated values have been examined and explained. This part of the verification can be found in Appendix 11. Output variables Finally the output variables are checked in order to complete the verification. After a warm-up period, when the process is in a steady-state, the number of requests that enter the system should be equal to the number that leaves the process. The results of two runs are listed in the table below. In Out 58472 58290 58424 58419 Difference -0.31% -0.01%

Run1 Run2

Table 2 - Verification output

18

Chile –

University of Technology

26 May 2005

6. Validation
A validation is done to see whether the model represents reality on the most important variables. There are different possibilities to validate a model. First there is the reality-check, the comparison between results provided by the model and the real values. In this case the check with reality happens by consulting several professionals with the results. The feedback they provide helps to shape and alter the model in a way it represents reality at its best. The results can be found in Appendix 12. Another way to validate a model is to check its responses on changes. This is very important, because the model is used to calculate (future) changes. It is not sufficient to represent the current situation, so in order to create a different situation it is necessary to give the model some external impulses. To provide some input variables with extreme values, the simulation has to respond according expectations. With an enormous increase in requests the utilization rate of the desk employees rises to 1, like that of the agents. The same effect, logically, occurs with a significant decrease of agents and desk employees. The expected effect of zero queues and low utilizations rates happens when the number of requests decreases to several per day. After a thorough walk-through with experts, in which the model logic as well as the outputs have been brought to attention, the simulation is validated and can be used to give conclusions about the present situation. It is ready to be used for alteration and implementing possible alternatives, helping in giving insight in probable situations in the future. End of the month boom During the evaluation of the model some new information was presented, which probably is of great influence on the subscription process. The changes regard the delay for a certain period of the requests. A corrupt bonus system causes this delay in the beginning of the process, just before the requests are passed on to revision by the supervisor. Agents have to make a certain amount of sales per month otherwise they do not get their bonus. The bonus is not raised if the agents make more sales. The result of this is that the sales agents will not submit a significant part of the requests until the end of the period in which they have to submit their quota. They save requests for agents that made too few sales in the end of the period, hence cover each others back. Because of this, in the end of the period, the final week of the month, the number of requests flowing into the rest of the system is about twice as much as in the previous weeks. The first three weeks about 33% of the requests are hold back by the agents and released the final week. The effect on the average utilization is nil, because the same amount of requests flow through the processes. The average subscription time increases rapidly, compared to the original results, with about 20%. The results will not 19

Chile –

University of Technology

26 May 2005

be taken into account in the analysis of future underwriting, for two reasons:   Uncertainty of data, distribution of delayed requests and procedures concerning this remuneration system. Probability that such a system will change soon.

Especially the first reason is important, because without sufficient data and knowledge about the exact procedures, it is very hard to estimate the real influence of it. It can have a significant effect on the way the simulation is set up as well, which makes the results even less valid. Nonetheless it is important to mention the possible effects of this system. This important issue will be brought up again in the recommendations, but the analysis of future alternatives will be based on a subscription process without this (ridiculous) bonus system. Assumed is that the requests are distributed evenly over a monthly period, that there is no difference between the first and the last week of a period qua number of requests flowing through the system.

20

Chile –

University of Technology

26 May 2005

7. Results simulation model
In the previous chapter the results of the simulation were compared to estimations given by experts. After some alterations in times which were uncertain the model has been made more representative. As mentioned earlier, the final results of the current situation are based on a simulation model without the system of bonuses that causes longer queues at the end of each month. Because of the lack of sufficient and validated information concerning this matter, the chosen approach regards a normal and consistent flow of requests throughout the month. It provides realistic output, on which future scenarios can be judged. The following table shows the distribution of the initial selling of policies. Most of it is done by agents calling clients and a small percentage by clients visiting an ING office. Front desk 5% Agent / iPAQ 95%

Initial selling of policy

Table 3 – Current situation

From now on, only the performance indicators will be mentioned. These are the most important variables, chosen by ING, which demonstrate how well certain alternatives or scenarios can be. The following results are calculated by the simulation model. Average subscription time Utilization Front desk employee Utilization Sales agent Utilization Sales manager Utilization Controller (TA) Utilization Controller (CP) 118 hours 2.5 % 58.9 % 18.1 % 67.6 % 81.4 %

Table 4 - Performance current situation

The utilization of the manager and front desk employee are quite low. This is very plausible, because most of the processes of these resources are outside of the system boundaries, like helping clients with questions by front office employees or having a meeting by the manager. There is not much to say about these results, because they are more or less nothing new. These results will be used as benchmark values for the scenarios described in the next chapter. The outcomes of each different scenario do not have to predict the future precisely, but the differences in outcomes of the variables (with the benchmark values) demonstrate quite accurate the potential of a certain alternative.

21

Chile –

University of Technology

26 May 2005

Future underwriting
In the future ING will make some big changes in the current underwriting process, which is described in the previous part. This part of the report can be considered as the most important part, because it focuses on these changes. As said, ING is interested in the outcomes of possibilities that will announce in the future or are already bound to be implemented. Before the end of this year, a so called ´Evaluation module (EM)´ (Módulo de Evaluador) will be implemented. The Business Rules Engine project (BRE) will run simultaneously with the implementation of the evaluation module. At the moment, it is not known exactly when the implementation of BRE will take place, but it will take at least one year. This part of the report serves to explain these (and other) possibilities, and present the results. In the first 2 chapter of this part, chapter 8, the different possibilities will be described and subsequently in chapter 9 scenarios will be created using these possibilities. The last chapters will deal with the results for each of the different scenarios and the recommendations.

8. Computerization options
The following part explains the possibilities in digitalizing the underwriting process in more detail. The options are put in a specific order, which roughly show the digital advancements. In appendices 13 until 16 are the specific processes that change for each option, including the different assumptions that have been done regarding the digitalization. Evaluation module At the end of the year 2005 ING is going to implement a program to computerize some decisions within the underwriting process and reduce the manual activity for executing these. This program is called ‘Evaluation module’ (Módulo de Evaluador) and is part of the general BRE. Because it could be interesting to observe the influence of the evaluation module exclusively and because it is to be implemented pretty soon, it is analyzed as a separate option. The Evaluation Module (EM) will digitalize between 50 and 70% of the first categorization processes. Business Rules Engine Another computerization project that just started at ING is the BRE-project. The objective of this project is to computerize a big part of the Technical Analysis (see appendix 4 Process Technical Analysis). At this moment, three full-time controllers are analyzing all the incoming requests. Estimated is that the BRE will take over 80% of their work. Digitalize all the paperwork within the Sales division Another option to computerize the underwriting process is to eliminate all the paperwork. This option will exclude some processes and enhance communication between Sales managers and agents. Within this option the 22

Chile –

University of Technology

26 May 2005

communication with the client will still be done with the necessary paperwork, because otherwise some other changes have to be made, which will be discussed later. Interaction with client We have seen different degrees of digitalization, ending with the most advanced alternative, i.e. the digitalization of all processes within the sales department. Still, this option can be extended even more with newer ways of communication with clients. These newer ways of communication will enhance the speed of transaction even more and will improve data storage. All the previous changes in the processes are assumed implemented in order to make these ways of communication possible. The following text will discuss proposals to change interactions between agents and clients. In the underwriting process the agent has to visit the client several times. These visits take a lot of time and therefore some other possibilities are proposed. These possibilities are only useful in a fully digital environment. The following regards three different possibilities of subscribing. As said, the previous digitalization is assumed implemented, but this is not all. For these new ways of client communication, a subscription program has to be written, taking over all the client’s documents and subscription papers. This program is compatible with normal web browsers and with iPAQs2. Agent with iPAQ Agents currently carry forms and documents when visiting clients. With this option the objective is to eliminate as well the paperwork in these processes as changing the information exchange from paper to digital. The agent will be carrying an iPAQ and uses this device to communicate directly with a server at ING. Therefore it is not necessary to visit the client as many times as he does at the moment. The idea is that it is possible for an agent, equipped with an iPAQ, to do the whole first part of the subscription process within one single visit. In this case, the request has to be simple and analyzable by the Evaluation Module and/or the Business Rule Engine. If the request is complex, hence needs extra background or a manual check, the agent might have to return another day to sign or get background. The processes that change are explained in the appendices. Client comes at the front desk The client comes to the office to do a request. The front desk employee digitalizes all information directly using a computer and possibly a scanner. All the possible following interactions with the client are done using an iPAQ, which will be explained now. Web based application
2

An IPAQ is a handheld computer and is chosen by ING as an example.

23

Chile –

University of Technology

26 May 2005

In this last option a different solution is being proposed for the interaction between agents and clients. The idea is to launch a website where clients can request a policy. This can obviously be done at home and there is no need for a agent to visit the client or for the client to go to ING. The rest of the process can be done by an agent using an iPAQ or client using the internet.

24

Chile –

University of Technology

26 May 2005

9. Options and results
There are several possibilities to analyze the options that are mentioned in the previous chapter. The chosen approach will deal with some of the digitalization options separately, but as well with some realistic combinations of them. Of every option and scenario the effects on the resources and average subscription time will be judged. The results of the current situation are shown in every table. The different options will be judged by comparing their respective results with these. In this chapter the first three digitalization options will be analyzed separately, the following chapter will bring the analysis of all the options within certain possible future scenarios. Evaluation module As explained earlier, the evaluation module is a digitalization alternative that is to be implemented very soon. This alternative will be implemented as a separate solution at first, and later it will be part of the more elaborate digitalization alternatives. The results of the Evaluation Module, only a small change in the whole subscription process, because it only partly affects two small processes, show a small impact on the performance indicators. Performance indicator Average subscription time Utilization Front desk employee Utilization Sales agent Utilization Sales manager Utilization Controller (TA) Utilization Controller (CP) Current 118 hours 2.5 % 58.9 % 18.1 % 67.6 % 81.4 % EM 50 % 119 hours 2.5 % 58.8 % 17.4 % 67.3 % 81.6 % EM 70 % 118 hours 2.5 % 58.9 % 17.1 % 67.4 % 81.4 %

Table 5 - Results Evaluation Module

The Evaluation Module regards two small processes done by the Sales manager. These processes confiscate only a few minutes of his time. As expected, the module will bring only a slight relieve for the manager, 1 %, corresponding with 1 manager fewer needed for these processes. The effect on other variables is negligible, due to the very small time profits. With regards to the subscription time or the utilization of the resources, this alternative will not have a very positive influence. The Evaluation Module is a part of the BRE and will be assumed implemented, with the computerization percentage of 70 %. Business Rule Engine An alternative that should have more effect, as well needs some more preparation, is the Business Rules Engine. This software program, together with a large database (filled with rules), will take over about 80% of the 25

Chile –

University of Technology

26 May 2005

technical analysis. The impact this has can be seen below. Performance indicator Average subscription time Utilization Front desk employee Utilization Sales agent Utilization Sales manager Utilization Controller (TA) Utilization Controller (CP) Current 118 hours 2.5 % 58.9 % 18.1 % 67.6 % 81.4 % BRE 119 hours 2.5 % 58.8 % 17.1 % 13.8 % 81.3 %

Table 6 - Results Business Rule Engine

The technical analysis is currently done by three controllers. With a Business Rule Engine that takes over 80 % of the analysis, not more than one controller is needed. The influence on the manager stays, because the Evaluation Module is part of the Business Rule Engine. Sales digital This influential alternative is already very advanced and can’t be implemented within a short time. But because it is technologically possible, it can be very interesting to look at the results of a large solution as this one. As said in the previous chapter, the paperwork within the sales division is replaced with software and digitalized information. Many of the flows and processes remain the same, because the interaction with the client happens as before, with the necessary paperwork. This situation can be seen as a transition phase, where later on the following scenarios can complete the technical environment. Before we analyze the future situations first the results of this alternative will be presented, because this semi-digital process is pretty much independent of regulatory advancements as digital signing and all legal documents remain the same. After this, the alternatives which need regulatory changes, will be discussed and analyzed. Performance indicator Average subscription time Utilization Front desk employee Utilization Sales agent Utilization Sales manager Utilization Controller (TA) Utilization Controller (CP)
Table 7 - Results Sales digital

Current 118 hours 2.5 % 58.9 % 18.1 % 67.6 % 81.4 %

Sales digital 83 hours 1.1 % 52.1 % 5.3 % 13.7 % --

Difference 35 hours 28 FD 66 agents 67 managers 2 TA 5 CP

The results of this intervention are clearly much larger than the previous digitalization alternatives. First of all the average subscription time for accepted requests decreases with more than 30% to average 8 working days, which is almost 2 weeks. Since there are much less processes for the resources to be done, unsurprisingly their utilization rates have shrunk. There are no people needed for control production, because the process it 26

Chile –

University of Technology

26 May 2005

totally digital. The other resources still belong to the subscription process, but their utilization is lower than before. This can be translated into a decrease of resources, as can be seen in the table. Of course, the decrease of resources regards the number of resources that will not be necessary anymore for the process of subscription, but they could be useful in fulfilling other processes. Next to the results shown in the table, there are huge profits in data storage, because almost all the information is stored digitally, which makes it much easier (and faster!) accessible and a lot less space-consuming. This alternative, which is still an alternative transitioning into a more sophisticated digital environment, brings great advantages, which are in:    Average subscription time Data storage Costs of resources

27

Chile –

University of Technology

26 May 2005

10.Scenario analysis
This chapter will deal with all the possible digitalization options, which will be implemented differently depending of the properties of each scenario. First is explained how the scenarios are obtained, then the scenarios and the results are written down and finally these results are judged. Interactions The following scenarios will represent a technologically possible and realistic future for the subscription process at ING. Apart from the previous digitalization, some other possibilities are implemented in this scenario, which deal with the client interaction. There are three ways of interactions possible:   At the front desk, with a front desk employee helping the client with the subscription. At the client’s house, with an agent, equipped with iPAQ, who takes information in digital format. Besides some brochures and possible background information, there will be no paperwork involved. The client using his or her internet connection to do a request, to accept restrictions or to make a (restricted) agreement.

Finally there is of course the agent calling the client to sell a policy, but this is not named above, because it regards only selling and not really an information exchange. The scenarios that will be analyzed will represent situations where all of these possible interactions are present, but with different percentages. Identified are four different moments of interaction:  Initial o o o selling of a policy, can happen in three different ways: Agent calling a client and visiting afterwards. Client visiting an office and buying a policy there. Client visiting ING website and subscribe there without help of any resource. Signing restrictions, three different ways: o Agent visits client, or still is at client (depending on complexity request) with an iPAQ  digital signing. o Client, still at the office, gets the restrictions from the front desk employee. o Client, still behind the computer, accepts (or not) the restrictions online. Deliver background o This is still done manually, because it regards official medical documents, which the agent has to digitalize at the office. Sign documents, this is the signing of the restricted contract, which can be done in three different ways: o Agent, possibly still at client, lets the client sign digitally with iPAQ o Client, still at front desk, signs electronically. 28

Chile – o

University of Technology

26 May 2005

Client signs electronically over the internet.

External variables Already a few possibilities for the ING to influence the underwriting process are defined, such as the Evaluation Module, Business Rule Engine and the agent with an iPAQ. However next to these possible new implementations, there are also some external variables in the underwriting process that can affect the outcomes considerably. Some of these variables cannot be affected in any way by ING, like the behavior of other parties within the underwriting process, for example the clients or the government,. Therefore it is important to take some of the most important influences into consideration when discussing future situations. A structural method has been used to define the three most important external variables, which can be found in appendix 18. The results of this short analysis are the following three external variables: 1. Personal contact Personal contact is a strong value for ING nowadays. Interaction between client and sales agent will always happen with the same sales agent. The client will be able to build up a more personal relation with ING in this way. It also is common-knowledge that elderly people prefer a more personal contact with a company. However these things can change. For instance, if we look at the Netherlands, things are already quite different. A lot of people like to decide things on their own, they do not like to be bothered or rushed by some sales agent who selling a policy. This external variable takes a possible change in attitude of clients into consideration. This variable considers a preference for personal contact, like now or even stronger, or the opposite, a more independent client, who prefers to decide on its own and does not need personal contact with any representative of a company. 2. Use of Internet This is a very general development that can be and already is, of very high importance to the business industry. The development of internet use has already developed itself tremendously and is still far from what can be realized. With the likelihood of the increase in Internet use in our mind, the future scenarios have to adopt this external variable as well. Increase in Internet use, but also Internet-related software, will lead to a lot of new opportunities to do business and sell policies. 3. Acceptance of digital agreements This last external variable will take the influence of the Chilean government into account. ING is bounded to follow certain procedures, concerning for example the archiving of client contracts. The Chilean government has certain rules, which ING has to follow and is therefore restricted in its possibilities with legal agreements. This external variable takes the possible policy direction of the Chilean government into account regarding the acceptation of digital legal agreements, such as digital signing and digital storage of contracts. 29

Chile –

University of Technology

26 May 2005

The three external variables are considered independent developments and can therefore be shown in a three dimensional space, which is divided in 8 quadrants. Each point in the graph represents a different future situation, but not every scenario is realistic. The future scenarios will be based on this graph and chosen on their credibility.
Use of Internet + -

Acceptance digital agreements +

Personal contact

+ -

Figure 1 - External variables

There are many different possibilities thinkable, but of course just a few will be analyzed. For every interaction moment, for instance, there is a certain percentage for each possibility (iPAQ, Internet, Front desk) within this moment. For example the initial signing, we could assume that 10% happens at the office, 30% over the internet and the rest on the old-fashioned way of calling a client and visiting. These initial percentages have of course influence on the other chosen percentages. The percentage of clients that use internet to accept restrictions will be less than the original percentage, because not every request is accepted. The following scenarios regarding these interactions have been written. For the Evaluation module a percentage of 70% is taken and for the Business Rule Engine a percentage of 80%. Scenario 1 – The friendly sales agent Use Internet: -1 Acceptance digital agreements: -1 Personal contact: +1 This scenario excludes the possibility of a total application via the Internet, because of problems with the legal system regarding digital legal agreements. Therefore the main focus lies on the Agent who, in contrary to the current situation, carries an iPAQ and visits the clients. 30

Chile –

University of Technology

26 May 2005

The initial request is done partly by clients visiting the front desk and partly by agents selling policies by phone. These internet applications regard only applications in this scenario, no follow-up actions as sign restrictions or documents can be done in this way. Agents with iPAQs are still needed to do these processes. Of course, if the client is at the front desk for an application and the request is simple (it does not need any manual interference and is processed digitally within a few seconds) these processes are done by the front desk employee. If the request is complex and needs manual processing, the client leaves and will be visited by an agent later on. This scenario is chosen from the perspective that legal arrangements still cannot be done over the internet. Because of this, the signing of the documents, the agent is still needed. The agent plays as well a significant and more personal role in building relations with clients. From the point of view that regulators move slowly with regards to changes, another important assumption is made within this context. The client still receives a contract at contract signing, which has to be picked up by the agent, when the request is not approved later on in the process. For the other scenarios is assumed that the client will only receive information about the policy, this will not have legal value. The following values have been taken for the initial way of selling a policy. As can be seen only a small part of the agents work of selling has been taken over by people (clients) using the internet. Front desk 5% Agent / iPAQ 80% Internet 15%3

Initial selling of policy

Table 8 – The friendly sales agent

The results of this scenario, of which internet is used only to do the initial request, most of the processes are digitalized and agents use iPAQs for client interactions, are in the table below. In this table and also the following tables are results shown of reductions in resource utilizations and eventually in a difference in how many resources are needed in the future situation. Some resources (sales agents and controllers) are therefore not necessary anymore. But for the other ones (front desk employee and sales managers) this means only that this amount of resources is not necessary for the underwriting process of the health insurances, but that they are necessary for other work. The underwriting process of the health insurances for the front desk employees and the sales managers is only a part of their total work. This has to be kept in mind when looking at the results. Performance indicator Average subscription time
3

Current 118 hours

Scenario 1 60 hours

Difference 58 hours

As said before, this part of the initial selling of a policy only regards the application. Further possibilities of interaction between a client and a sales agent (sign restrictions, documents…) are not incorporated in this future scenario.

31

Chile – Utilization Utilization Utilization Utilization Utilization

University of Technology 2.5 % 58.9 % 18.1 % 67.6 % 81.4 % 1.8 % 38.6 % 5.1 % 13.7 % --

26 May 2005 14 FD 200 agents 70 managers 2 TA 5 CP

Front desk employee Sales agent Sales manager Controller (TA) Controller (CP)

Table 9 - Results scenario 1

In a semi-digital environment which is created with this scenario, all the previous digitalization options have been implemented and a small part of the initial applications happens over the Internet. Agents still visit clients, now with iPAQ, and clients still receive their contracts. The average subscription time in this scenario has been halved, the time now until a request has been accepted and send by mail. For the resources the implications are even bigger. As seen before, there is no need anymore for Control production in a digital environment as this one is. This change, and other changes that already have been discussed, will not be handled again in this scenario analysis, only new things will be dealt with. At the resources the biggest changes occur at the utilizations of agent and manager, which decrease dramatically. In this scenario and with the described automation computerizations implemented, the process can be done with a lot less of these resources (in total about 135 less). Scenario 2 – A bit of everything Use Internet: 0 Acceptance digital agreements: +1 Personal contact: 0 The digital legal agreements are no problem anymore in this scenario. Therefore the two main possibilities of applying are through the sales agent and Internet, next to the visits at the front desk. This scenario is pretty much like the first one, except for the clients that apply through internet. These will finish their application, and possibly accept restrictions, in the same way, except of course when the request is complex and needs manual analysis. If a simple request, there is no need anymore for an agent to interfere in this process. However, still a large part of the requests will be done by the agents in this scenario. The main focus of this scenario is the level of personal contact. Because of less interest in personal contact the Internet option to apply for a policy becomes more attractive. Because the Internet application will lead to fewer resources needed, ING could give a slight discount for example to clients who choose for Internet applying instead of the other options. Therefore their share in this scenario is much larger. Front desk 5% Agent / iPAQ 55% Internet 40%

Initial selling of policy

Table 10 – A bit of everything

32

Chile –

University of Technology

26 May 2005

The results for this option, with a significant part of the processes being taken over by the clients themselves, are in the following table. Performance indicator Average subscription time Utilization Front desk employee Utilization Sales agent Utilization Sales manager Utilization Controller (TA) Utilization Controller (CP)
Table 11 - Results scenario 2

Current 118 hours 2.5 % 58.9 % 18.1 % 67.6 % 81.4 %

Scenario 2 58 hours 1.8 % 31.2 % 5.1 % 13.2 % --

Difference 60 hours 14 FD 272 agents 70 managers 2 TA 5 CP

For this scenario you would expect a smaller average subscription time, because the application time through internet is much smaller than the agent visiting clients. Still there is not much difference between the scenarios. This is because assumed is that the agents immediately go to the client for a request, instead of making an appointment, which can take weeks. The time of a client doing an application over the internet is just a little smaller than one through the ‘traditional’ way, with an agent. The only big difference is again in the resources, this time the utilization of the agents shrinks even more, as you would expect. Because more people taking over processes that are usually done by agents and agents do not have to visit clients that often anymore, 159 agents less are needed. The use of internet increases day by day and in the following scenario a large role will be allocated to this upcoming and promising medium, which can cause even a smaller need for agents. Scenario 3 – ING.CL Use Internet +1 Acceptance digital agreements +1 Personal contact -1 In this scenario the focus is on the internet possibilities of selling insurance policies through the Internet. Agent who visit clients and clients who apply at the front desk are a minor percentage of the total applications. Like the previous scenario, there are no problems with digital legal agreements. This scenario is chosen from the point of view that most of the agent’s work can be done by the client itself. A clear and fast website has to make sure that without any problems future clients can finish their application and do their request fast and easy. For clients there will be advantages of doing it over the internet, such as the freedom of canceling at any moment (which could be disadvantageous for ING) and the possibility to do the request at any time. Applying over the internet can be a disadvantage as well for clients who need help in filling out the forms. Applications over the internet are of course much cheaper for ING than applications which have agents involved. Therefore clients who do everything 33

Chile –

University of Technology

26 May 2005

over the internet, have lower initial costs and could be awarded for this with for example a discount (same as in the second scenario). The value of internet applications will be dealt with below in the results. This scenario will have twice as much internet applications as the previous, so an easy comparison can be made. Front desk 5% Agent / iPAQ 15% Internet 80%

Initial selling of policy
Table 12 – ING.CL

The results of this scenario are below. Performance indicator Average subscription time Utilization Front desk employee Utilization Sales agent Utilization Sales manager Utilization Controller (TA) Utilization Controller (CP)
Table 13 - Results Scenario 3

Current 118 hours 2.5 % 58.9 % 18.1 % 67.6 % 81.4 %

Scenario 3 57 hours 1.7 % 19.9 % 5.0 % 13.3 % --

Difference 61 hours 16 FD 383 agents 70 managers 2 TA 5 CP

As predicted, the number of agents needed in this futuristic scenario shrinks even more and falls to one-third of the original number needed. The average subscription time still remains at around 60 hours for this pretty efficient system. Still, this is not the time a client has to sit behind the computer, because the conditioned contract is done at Sign documents, which is still in part 1. The following processes deal with notification and controlling the request, but the client does not need to sit behind the computer for this, it is all done without a client interaction. In this way a request over the internet can be done within half an hour.

34

Chile –

University of Technology

26 May 2005

11.Summary results
This chapter will be handling the most important results from the previous two chapters. First an overview of the most options and scenarios can be seen in table 14. Option/scenario Evaluation Module BRE Sales department digital Characteristics and prerequisites The Evaluation Module takes over a part (70%) of the decision process at the beginning of the subscription process. The Business Rule Engine project aims for the digitalization of Technical Analysis, another decision process. It will probably automate 80% of this process. This possibility regards the total digitalization of paper flows within the sales department. Contracts to clients etc. are still handed in the normal way, but every document coming into the sales department is stored digitally, which makes some processes redundant. Besides the sales department, the agent-client interaction has been digitalized as well, with agents using iPAQs connected to and ING-server. In this way many processes need only one visit of the agent. Clients can do their initial application over the internet. This scenario is mostly independent of outside factors as governmental changes and availability of internet. This second scenario assumes that clients can do the whole application over the internet and that digital signing and restricted contracts are allowed governmentally. Most of the client-interactions happen over the internet and by the agent. Scenario three focuses on clients using the internet as the most important way of doing requests. Agents and front desk employees take care of only a small part of the interactions and processes with clients.

Scenario 1

Scenario 2

Scenario 3

Table 14 - Overview options and scenarios

Of every option and scenario, the most relevant results are explained in the table below. In italic under every option/scenario the performance indicator is written on which the option/scenario has effect. Option/scenario Evaluation Module BRE
- Controller Technical Analysis

Results The Evaluation Module shows little effect on any of the performance indicators. The Business Rule engine itself cannot bring much change in process times or resource utilities (except for Technical Analysis) either, but it has a lot of possibilities for future applications. Besides, it brings all the advantages of the option itself, such as better and faster analysis, easier and better Rule management etc. 35

Chile – Sales department digital
- agents, managers, control production - subscription time

University of Technology

26 May 2005

Scenario 1
- agents - subscription time

Scenario 2
- agents

Scenario 3
- agents

This option is not realizable within a short time. Many processes have to change in order to create a digital environment like this. The profits that can be gained with this opportunity one the other hand, are great too, with a much smaller amount of resources needed (40 agents, 12 managers and 5 controllers CP less). As well the average subscription time decreases with 36 hours, which is about 4 working days. In this scenario, where the sales agent still plays an important part, the digitalization has focused not only on the processes within the Sales Department, but also on the interaction with clients. Part of the initial subscription is done over the internet now and all the client-agent interactions happen with an iPAQ, which is connected to an ING-server. Because of a significant smaller travel time (less visits), the subscription time and the number of agents needed shrink again. In this scenario more people do all their subscription interactions over the internet. Obviously less agents are then necessary, because some of the processes itself are being taken over by the clients themselves and the clients do not have to be visited by agents. This causes another big drop in the utilization of agents. This final scenario doubles the amount of subscribers over the internet and the role of the agent his of less importance. The expected result of an enormous decrease of needed agents happens and a bit more than half of the agents of the current situation are needed. The influence on the average subscription time does not change that much within the different scenarios, all around 60 hours (6 working days). The client can do an application over the internet in half an hour.

Table 15 - Results overview options and scenarios

36

Chile –

University of Technology

26 May 2005

12.Conclusions and recommendations
The underwriting at ING, at least for the subscription process at product division Healthcare, could happen much more efficient. ING is making a lot of effort in reshaping processes and implementing systems that have to improve the overall process.  The systems that will be implemented within a few years are the Evaluation Module and Business Rule Engine. Except for an improved data storage and rules management system, these digitalization options will not have a big effect on the resources in the subscription process, neither on the average subscription time. A much broader digitalization of the Sales division, including EM and BRE, has more effect, but needs investments and a lot of analysis. This option will bring large effects, not only on the average subscription time, but also on the resources. The process will get faster and will be done more efficiently. The data storage and rules management are as well improved, which has many positive side-effects. Technological possibilities and several external variables have been used to generate scenarios. With a progressive government the possibilities for subscription through internet are the best, as well are the benefits in number of resources needed and the average subscription time. A client can do a request solo and within half an hour over the internet. Agents visit clients equipped with iPAQs connected to an ING-server. Paperwork belongs to the past. With a less progressive government and less people on the internet, the profits gained by digitalizing every possible thing are still great.

With regards to the results there are several recommendations to be done. ING needs to focus on a total digitalization of the subscription process, including better data storage and rules management system. Evaluation Module and Business Rule Engines will not bring great advantages in cost reduction or a decrease in subscription time. Therefore these systems are only a step towards a more advanced technological system, which has to make the system faster, clearer and better manageable. These recommendations can be summed up as follows:  Implement BRE and EM, but see these options only as prerequisites to a more influential overall system.  Conduct research for a more representative model, with better validated data. Better/more information about: o Route times o Number of clients per day o Process times o Distributions o Logic  Focus on a more advanced digital system by analyzing all digitalization possibilities within this model.  Do costs-benefits analyses about these possibilities. 37

Chile – 

University of Technology

26 May 2005

Do scenario research towards external variables as use of internet, governmental change and client’s desires regarding subscription.

In Appendix 19 a small overview is given regarding possible further analyses, divided in a Process part, which deals with the enhancement of the simulation model, and a Scenario part, which deals with future possibilities. Besides above recommendations, a short remark has to be made on the bonus system as described in chapter 6, Validation. Obviously, this corrupt system is bad for the company and needs to be changed. A quick and thorough research is needed to investigate possible difficulties in changing the system or hidden advantages. Finally, one other recommendation can be made regarding continuation of this analysis. An important recommendation is to buy the software program Arena and let experts on the processes within ING make more representative models. Not only can they make models of the subscription of Healthcare, but any other process for any other product. Because they will be made in Spanish, they can be passed on to other analysts working for ING. This will bring a more insight into future reengineering possibilities for business processes within ING. Because it takes time to apprehend and use the software, it might be cheaper to let other interns, affiliated with the program, be part of the project. These students bring their software themselves, with a one-year license, which is cheaper than buying the program.

38