Saturday, July 3, 2010

WEEK 13 ( 28 JUNE- 2 JULY)

This week I did some testing on Microsoft Outlook. I learned how to fax using Microsoft Outlook. This was done through UNICEED account. Then, the testing was recorded step by step in manual form. 


Microsoft Outlook - Fax

I edited the FAQ for UNICEED. Besides that, I also did some settings in Superceed website for the call flow. Since there was error in the website, I was asked to upload the voice file again.

Finally, it was time for Beta testing. My colleagues and I pretended to be agents that answers the calls. Then , we answered the calls for different paths.That was my last task for Industrial Training.

Tuesday, June 29, 2010

WEEK 12 (21-25 JUNE)


My task for this week was fully involving the superceed/vcc/inbound website. I had to create path for each options and level in the website. I also had to upload the voice file that I created last week. Before doing this, I had to create certain categories in the website. 

Superceed website


One of the part of the call flow

Next, I had to create agents/supervisor to attend the calls for each path. I have created about 20 agents for all the paths in the call flow.  

Agents/supervisor list

Besides that, I also created 30 frequently asked question by customers and prepared the answers for each questions. I type out the Q&A and convert it to PDF file. This was then uploaded in the website under knowledge base section. This was done for the agents so that it will be easier for them to answer the customer's questions. 
 
FAQ in knowledge base

Finally was the attributes for agents. I had to divide it into categories and sub-categories.

For example:

Cat                                Subcat
Languages                    English
Department                   Sales & Marketing
Sub                               Features
Specialized on               Voice

Monday, June 28, 2010

WEEK 11 (14-18 JUNE)


This week I was asked to design a call flow for the company's product UNICEED. A call flow is an interactive voice response application. The call flow describes how the caller enters the application, the options and inputs (key presses) that are provided to the caller, and the application's response to these inputs. Before designing a call flow, certain criteria need to be planned. For example, how many languages are available for this customer service or what departments do I route the calls to. 

Call Flow Diagram

The call flow is divided into 4 levels and 3 options. When a customer call UNICEED customer service, firstly they have to choose their preferred language. Then, departments that they wanna deal with. Next would be the subcategory of the department and finally is what they wanna specialized on. 

There 3 options available are English, Bahasa Melayu and Mandarin (level 1). Each language are divided into sales & marketing, technical assistance, billing, general enquiries and automated guide. These are called the departments (level 2). 

Under sales & marketing department, there are features, packages, rates and reseller information. Whereas under the technical assistance, the choices available are voice, fax, email and sms. Payment, renewal and reload options are available under billing department. Other than that, if the customer choose general enquiries, it will directly connect them to the telephone operator. Under the automated guide are sales, technical and billing. All these are sub-categories (level 3).

The final level is level 4 (specialized on). This level is only available for sales & marketing and automated guide. Under sales-features, there are voice, fax, email and sms. Whereas under rates, available options are domestic and international. For reseller information, there are uniceed seller program and private label program. Finally, the under each sub-categories of automated guide, there are few frequently asked questions (Q&A). This option will assist the customer when it's after office hours. 

All these levels are then translated to Bahasa Melayu and Mandarin in order to cater customer's needs in other languages. Besides that, I had to convert those text to speech. This was done by using Text-To-Speech software. After recording the voices, I edited the voice file in wavepad editor by changing the sample frequencies, bit rate and many more. 






Monday, June 21, 2010

WEEK 10 (7-11 JUNE)


Beginning of this week, my colleagues and I had a meeting with our boss. We discussed about all our task that have been done for the past 2 months. In the meeting, I have been briefed about call routing strategies and virtual contact center. Besides that, I was also given new tasks for the week.

 In the meeting room

Nowadays, someone in England can make a phone call to Australia, or that someone in United States can read web pages that are on a computer in Canada. A large network is affected by many factors which are often hard to predict. There can be busy and quiet periods through the day. For example, if a television program has a phone-in vote, there can be a sudden overload at one point in the network. This is when routing calls comes in handy. These routing schemes work by searching out the spare capacity in the network so as to route calls away from parts of the network that are broken or full and into parts that are underloaded.




Call routing strategies is divided into:-
  • last agent
  • round robin
  • lowest total talk time
  • fewest received calls
Last agent

Route new calls to the general queue and repeat calls to the last agent contacted with overflow to the general queue. In this routing strategy, calls that queue for a specific agent are overflowed to the general queue after a pre-set period of time (i.e, wait, 1, 10, 20 and 30 seconds for a specific agent). Repeat callers waiting for the last agent have a higher priority than general callers.

 last agent

Round robin

If more than one agent is available, the system will try the agent who goes after the last one who took a call in a round robin fashion. Imagine 3 agents are in a circle. A call comes in, and the first agent takes it. Next time a call comes in, agent 2 takes it. Next time it will be agent 3, and then back to agent 1, as long as all of them are available. This routing method is appropriate for situations when most calls are of comparable length and all agents need to take a comparable number of calls. In other words, it balances the load of incoming calls based on number of calls.

I also learned about contact center.

Virtual Call Centers

Skills Based Routing can be implemented as part of a much more elaborate call center scenario, the virtual call center. Several vendors are now providing fresh looks at how a virtual call center would operate within a network environment. The application sits on top of Centrex, central office or PBX environments, allowing agents from multiple geographic locations to log into a single virtual ACD environment. Think of it as an ACD without the equipment, but totally dependent on the public switched network. In this scenario, routing calls based on agent skills is a network requirement and easily implemented. Calls are labeled as a primary skill or a non-primary skill. Matching calls to agents becomes easier and less costly within the network and easier for the customer and less costly for the call center.


Vendor Approaches

There are a variety of companies providing products and services with special features focused on Skills Based Routing.

Blended Call Centers

Blended environments are call centers where inbound calls and outbound calls are handled by the same agents. Inbound call activities are assigned as one skill and outbound call activity is assigned to a second skill. Vendors supplying call routing systems for this environment may include enhanced ACD and IVR functions with CTI features.

Premise-Based Contact Center
Call center have been built on PBX equipment that is owned and hosted by the call center operator. The PBX might provide functions such as Automatic Call Distribution, Interactive Voice Response, and skills-based routing. The call center operator would be responsible for the maintenance of the equipment and necessary software upgrades as released by the vendor.

Lastly, I learned call queue and call distribution strategy.
  • Call Queue is to offer customer the best waiting time possible.
  • Call Distribution Strategy : supervise talk time (average 3 min) and supervise the etiquette of agents.

    Friday, June 4, 2010

    WEEK 9 (31 MAY- 4 JUNE)


    Finally, it's the last month of Industrial Training. Yahoo!!! Last month of the internship I'm gonna be doing beta testing. So, this week was preparation for the beta testing in order to participate in real contact center environment.

    Beginning of the week, I was asked to write a call script. A call script consists of Q&A designed anticipate questions from the caller, and what answers you could provide to each of the anticipated question. I was assigned to Syarikat Bekalan Air Selangor (SYABAS). This call script contains of 30 Q&A's in General Enquiries, Customer Service and Billing.

    End of the week, I compiled a list of 50 contacts in Mid Valley. The contacts are from companies in South Tower, North Tower and Boulevard. After that, I did an additional 30 contacts for KL Sentral companies. This was done in order to test the inbound and outbound calls. 

    WEEK 8 (24-28 MAY)


    Erlang algorithm is a common algorithm used in Workforce prediction in a contact center environment. This week I learnt about Erlang. Basically, erlang is a unit of traffic density in a telecommunications system. One erlang is the equivalent of one call (including call attempts and holding time) in a specific channel for 3600 seconds in an hour. The 3600 seconds need not be, and generally are not, in a contiguous block. Erlang is divided into 3 which are Erlang A, Erlang B and Erlang C.

    In call centers, Erlang A is used, to support solutions of the staffing problem, such as how many agents should be answering calls during a specified time period. In order to apply Erlang A, it is necessary to input values for its four parameters: λ, μ, θ and n. Parameters λ and μ are calculated for every hourly interval. We also calculate each hour’s average number of agents n. Because the resulting n’s need not be integral, we apply a continuous extrapolation of the Erlang-A formulae, obtained from relationships developed in. Finally, for θ we use the formula P{Ab} = _ • E[W].


    A traffic engineering model that assumes that an offered call is cleared immediately, with no queuing. In other words, Erlang B assumes that a call encountering blockage will not appear again. Either the caller will hang up and not attempt to place the call again, or the call will automatically be routed over another circuit if one exists, even if the use of that circuit is more expensive. 

    Erlang B is the formula to use when a blocked call is really blocked. For example, when somebody calls your phone number and gets a busy signal or tries to access a tie trunk and finds it in use. It is built around three variables which are Servers, Traffic, and Grade of Service.

    The most common traffic engineering problem involves sizing a trunk group. For example, how many trunks are needed to carry your toll-free calls, how many tie trunks between two offices, how many ports into your voice mail system, or some similar question. Erlang B handles that relatively easily, in four steps:

    1. Collect traffic data
    2. Determine the Average Busy Hour
    3. Choose a target Grade of Service
    4. Use Erlang B

    Example of Erlang B


    Erlang B Formula 


    Erlang C formula is used when a blocked call is delayed. For example, when someone calls your call centre and must wait for an agent to take the call. It uses the same three variables, plus the average length of each call, to calculate the probability of being delayed and how long the delay is likely to be.

    Because Erlang B is so simple to use (insert two numbers, it calculates the third), many people assume that Erlang C will be similarly easy. But it is not true, even basic Erlang C calculations are difficult, and more complex ones can be daunting indeed. Erlang C is most commonly used to calculate how long callers will have to wait before being connected to a human in a call centre or similar situation.

    Erlang C Formula

    After learning about Erlangs, I searched for erlang sample excel sheet calculations and formula to test it out. Firstly, I searched Erlang B.

    Example of Excel Function for Erlang B:
    ErlbBlockage(nsrv,trafficInErlangs) returns the Erlang B blockage for a specified number of servers and specified offered traffic. The nsrv is the number of servers (can be any non-negative number) and traffic in Erlangs is the offered traffic in erlangs (can be any non-negative number). For example, =ErlbBlockage(3,10.4) returns Erlang B blockage for 3 servers and 10.4 erlangs of offered traffic. This then returns the value 0.7411421.

    Besides that, there are more functions for Erlang B in excel such as ErlbNsrvFromBlockage and ErlbTrafFromBlockage.

    Example of Excel Function for Erlang C:
    ErlcFractionDelayed(nsrv, trafficInErlangs) returns the probability that a customer arriving at the queuing facility will experience a delay before beginning service. For example, =ErlcFractionDelayed(11,10.1) returns the probability that a customer will experience a wait, when there are 11 servers and 10.1 erlangs of offered traffic. This then returns the value 0.71095477.

    Besides that, there are more functions for Erlang C in excel such as ErlcFractionOk, ErlcNsrvFromFractionOk,  ErlcWait and ErlcNsrvFromWait.


    Saturday, May 29, 2010

    WEEK 7 (17-21 MAY)


    This week task looked simple but it's actually very hard. I had to find out Text-to-Speech plugin asterisk suppliers. I also had break it into 4 sub-topic such as best virtual person, integration with vendor products, integration with custom 3rd party application and product summary.

    Suppliers that I manage to find are as listed below:
    • The Festvox Project
    • Cepstral
    • Acapela
    • Power Text to Speech Reader 2.31
    • Sayvoice Text to Speech Reader
    • Word Talk
    • ACE-HIGH
    • Sayz Me
    • NextUp Talker
    • Loquendo
    Besides that, I also had to find out contact center vendors (premise-based and hosted).  Many contact centers are heavy users of communications technology. Premise-based are the traditional equipments and software that you need to install in the office. Hosted are those which you do not need to buy, just need to subscribe. 

    Below are the contact center vendors (premise-based):
    • Nobel Systems
    • Aspect
    • Cisco
    • Computer Talk.Tech
    • Upstream Works
    • Zeacom
    • Interactive Intelligents
    • LumenVox
    • TouchStar
    • Siebel
    Below are the contact  center vendors (hosted/on-demand):
    • Cincom
    • Contactual
    • Envox Worldwide
    • LivePerson
    • Smoothstone
    • TeleTech
    • UCN
    • Five9
    • Packet8