Call Center Agent example essay topic
Remember the 15-digit account number he entered? To Mycorp it's much more then 15-digits: it's about every product he has purchased using his Mycorp credit card; his language preference; if he is past due on his bill; if he is considered a high value customer to the organization; the type of card he has, and much more. The call center of today is evolving at an extraordinary rate. New technologies are popping up all over the globe bringing to the table new and exciting products that will help companies to service their customers in faster and smarter ways.
One of the products with the most impact on the market today being used to accomplish this is Intelligent Call Management. Many large organizations such as American Express, Wells Fargo, Charles Schwab, and Discover Card are using Intelligent Call Management to provide better service to their customers. A recent survey discovered that 90% of today's call centers are using voice-processing applications, while another 54% have introduced Intelligent Call Management into their call center. Intelligent Call Management, or ICM, cannot be accomplished with one single piece of hardware. It involves multiple peripherals working together, each providing a different functionality that contributes to the efficiency of the call delivery. ("Are You", 2004) Computers and the telephone have technologically collided and as a result CTI, or Computer Telephony Integration, was born.
The CTI component is the most important piece in an ICM environment. With CTI, companies are improving their efficiency, lowering their operating costs, and building the infrastructure for new business opportunities. Surveys show that 64.2% of call centers use CTI in their daily operations. CTI provides call centers the ability to connect their telephony peripherals to a computer system so that data messages or "events" can be communicated back and forth over a data network.
Events are the key to keeping track of all the calls coming in and out of the call center. For example, every time a call rings an agent's phone, a "CALL DELIVERED" event is generated and sent over the CTI connection. Every time an agent answers a call, a "CALL ESTABLISHED" event is generated and sent over the CTI connection. Many different types of events exist and act as triggers for anything that might be listening to the communication. Located with in the event's message set is a unique call identification number that is used to keep track of the call throughout the data network.
This Called is also used to match up the voice call with the data so that the two elements do not get separated while traveling between multiple peripherals. Located within the events message set are call variables. A call variable is a string of bytes set aside for the system to store information about the call or caller, and they travel with the call until an agent answers. The data elements stored in the call variables are used to make intelligent decisions about where to route the call.
For example, if a caller has a "Y" in call variable 1, byte position 1, it is identified as a High Value card member and should be treated differently than someone who has an "N" in the same byte position. Another example would be to identify the customer's language preference. If call variable 2, bytes 3, 4, and 5 contain "ENG", then Mycorp knows to send this call to an agent who speaks English. Virtually any piece of known information can be stored into the call variables associated with the call. Based on these values, the routing destination can change dynamically.
The CTI connection also allows the call routing infrastructure to monitor the status of the telecom peripherals on the network. Since all the peripherals have CTI connections, the call router can verify if a destination peripheral is online and available before making a decision to send a call to it. An example would be an emergency situation. For example, assume there was a power outage in the Texas call center. The call router has the ability to determine if the site is available by sending heartbeats down the CTI connection.
If a response is not returned with in a specific amount of time the call router will automatically re-route calls to a different call center. After having discussed the glue that holds all the telecom components together, now is a good time to introduce the rest of the components. The most commonly known component in the ICM or Intelligent Call Manager environment is the VRU, or Voice Response Unit. The VRU is a computer that runs voice applications and hardware that allows it to communicate with the caller. Its primary function is to answer calls and provide voice menus to callers that allow them to service themselves without agent interaction. In a credit card servicing environment, it is the peripheral that customers interact with that prompts them to enter a card number over the telephone.
In newer environments it is not uncommon for the VRU to be equipped with voice recognition hardware that allows the customer to navigate through the menus by speaking instead of pushing the buttons on a telephone. Also available is text-to-speech technology, which speaks dynamically created content back to the customer such as account balances, telephone numbers, and addresses. The hardware associated with a VRU usually consists of an operating system, UNIX in most cases, incoming telephone lines usually spanning 24 line intervals, and a CTI link connecting the VRU to other peripherals on the network. The ACD, or Automatic Call Distributor, is a telephone switch that sits at the front of a company's voice network. All calls arriving from the network carrier get delivered to the ACD using TDM, or Time Division Multiplexing. This is a process of taking multiple phone lines from the carrier and multiplexing the information into a single transmission where each frame is sent in specific time slots.
When then transmission reaches the destination site, the information is de-multiplexed back into individual phone lines. When a call arrives from the carrier, it arrives on what is called a VDN, or Vector Directory Number. A VDN is an extension akin to those that employees are used to dialing in an office environment. It usually consists of 4 to 6 digits, and is only connectable from inside the company's voice network. Every VDN is associated with a vector or multiple vectors.
Vectors are programming steps that allow the switch to logically manipulate the call. For example, a vector might have a step that plays an announcement, collects touch-tone digits, routes the call to another VDN, or most importantly queues the call to a skill group. Skill groups are agents that are grouped together because they have the same skill sets. When the ACD queues a call to a skill group, it places the caller on hold and waits for an agent to become available who has a specific skill set. When an agent in the skill group becomes available, the call is transferred and the agent services the customer. The ACD has CTI connectivity through hardware blades known as the Multi-Application Platform on Definity (MAPD) that allow it to communicate with other peripherals on the network.
MAPD is an Avaya related term but other switch manufacturers have similar devices that provide CTI connectivity. (Stern) The Central Controller is the brain of the entire operation. It is, in essence, the mother ship. The portrayal of the ICM environment applied here uses a software product from Cisco Systems called "ICM". Some individuals will refer to it as an "ICR", or Intelligent Call Router. The Cisco ICM software uses Microsoft Windows 2000 as the operating system and Microsoft SQL Server 2000 to house all routing logic and configurations.
Real-time and historical call routing data are also stored in the ICM database. The function of the Central Controller is to listen to all CTI communications and wait for key triggers. These triggers are called "Route Requests", and are sent by all peripherals on the network. When a peripheral is in a position to route a call, it sends a message down the CTI link to the Central Controller. In most ICM environments the route request tells the Central Controller what type of routing decision it needs to make. For example, a "Route Req 00" message is sent by the ACD and is interpreted by the controller as, "I have a caller coming in from the carrier; which VRU or Voice Response Unit would you like me to transfer this caller to?" A "Route Req 01" data message instructs the Central Controller to perform a database hit to obtain all the information it can about the caller and return the caller to the VRU to continue self-servicing.
A "Route Req 02" asks the Central Controller to make a routing decision based on the contents of the call variables. In other words, it is asking, "What agent do I send this caller to?"Route Req 01" and "Route Req 02" are performed by the VRU. (Rohde's 1997) When the central controller recognizes a route request coming in over the CTI connection, it executes a call routing script based on the type of route request it received. Call routing scripts are process flows that contain logic for processing calls. To paint a picture of what a call routing script looks like, it would be comparable to a Microsoft Visio document. The script contains many nodes that are connected by lines depicting your traditional programming functions of "if this then do this, else do that".
The call starts at the top of the script and works its way down to the bottom, passing through business rules. If the criteria located in the call variables meets a specific business rule programmed in the node, then ICM will return a "label" back to the peripheral that made the request. A label is little more than a response back to the peripheral instructing it to dial a VDN. Clearly the ICM Central Controller is an extremely important component because it makes all the routing decisions. It is for this reason that companies are able to run multiple Central Controllers in a duplex configuration.
Each Central Controller is an exact mirror of the other and they are connected together by an internal network strictly used to send and receive heartbeats. If the active Central Controller should crash, the Central Controller on standby will recognize this because of the loss of heartbeats being sent between the two. The standby Central Controller then will take control of the network by becoming and it will start to respond to all route requests. If for some reason both Central Controllers were to become inoperable, the peripheral that made the route request has fall back logic programmed to handle the calls. In the ACD, vector programming exists that will queue the call to a location if it times out making a route request. The VRU has what are known as "Default Routing Tables" that it uses to send the call in the event of a time out.
Now on to a virtual tour that demonstrates how all of the above components work together. A customer dials Mycorp's 800 number and the network carrier delivers the call to Mycorp's ACD. The ACD programming transfers the call to an internal VDN. When the call arrives, the VDN's vector programming is executed to make a route request (Route Req 0) over the CTI link to the Central Controller. The Central Controller executes the call routing script associated to "Route Req 0". In the script's logic the first thing checked is if both VRU systems are online.
If they are, which VRU is currently servicing the least amount of callers? For this example, the assumption is that VRU 1 has the lightest load. The Central Controller returns a VDN (dial able extension) back to the ACD. The ACD dials the VDN returned by ICM and the call is delivered to VRU 1.
The customer is now prompted by a voice application on VRU 1, which asks you for a language preference. "Hi, and thank you for calling Mycorp. Please enter 1 for English, and 2 for Spanish". The customer presses "1" on the touch-tone phone for English and is then prompted again, but this time for a 15 digit account number. After entering the account number, the VRU writes the account number and language prefix into the call variables associated with you call, and sends a route request (Route Req 01) over the CTI connection. The Central Controller receives this message and executes the routing script that is associated with Route Req 01.
In the script's logic, the first step is to make sure the ICM database is online and available. Next the ICM performs a database dip to a mainframe system to obtain customer information using the customer's account number as the key. ICM then populates the customer's information it has received from the database into the Call Variables and sends them back to the VRU. The VRU then proceeds to service the customer by communicating a menu of options that the customer can use to take care of business that does not require an agent's assistance. If for some reason the customer would like to talk to an agent, he or she would press "0", as in most traditional phone systems, and the VRU sends a route request (Route Req 02) message over the CTI connection. ICM receives the route request and executes the script associated with Route Req 02.
In the logic of the script, the ICM starts by looking in a specific byte location of the call variables to locate the product type, which was propagated by the earlier database hit. The next check looks to see if this is a High Value customer, meaning he or she spend quite a bit of money at Mycorp and needs to be sent to an experienced agent. The following check looks at the language prefix that was selected when prompted by the VRU. Now the ICM has almost identified enough information about the caller to make a routing decision that will provide him or her with the best possible customer experience. There is, however, still one important check that the ICM will perform that has a major impact on call center costs and customer satisfaction.
The ICM checks for agent availability to find the best agent for the required service in the least amount of time. ICM accomplishes this by using the CTI link to obtain agent skill information from the ACD. Once the ICM identifies a candidate, it returns a label back to the VRU. The VRU dials the label, which send sends the call to a VDN located on the ACD. The VDN points to a vector and the vector programming queues the caller to a specific skill set, which in turn allows the customer to be serviced by the appropriate agent. How does the agent know so much about the caller?
As stated before, the call variables that were constantly being updated by the VRU and ICM travel with the call everywhere it goes. When the call reaches an agent, so do the call variables. The agent has a computer on his or her desktop running a specialized application. The application has a CTI connection that is used to communicate with the Central Controller.
When the call arrives at the agent's phone, the central controller provides the call variables through the CTI connection. The desktop application reads the call variables and populates the agent's screen with all of the customer's information. At least 30 seconds can be taken off call time if the customer's information is populated on the agent's desktop. ("Are You", 2004) ICM is a great solution, however it does not come without global implications. The configuration explained at the beginning of this paper only depicted a single site.
Most call centers are discovering that it is cheaper to service customers from overseas than it is to service customers from within the United States. For ICM to be able to transfer calls to a remote location and still keep the call variables intact, the destination call center needs to be ICM enabled. This means that the destination's call center has to be able to talk to the central controllers through a CTI link. Most companies do not want to spend the money to build a call center overseas so they contract with an existing call center. Since not all call centers use the same hardware vendors and software versions, the destination site will require a piece of hardware called a PG or Peripheral Gateway. A PG is basically a translator.
It takes instructions or messages from one brand of ACD or VRU, and converts them into another format. In this case, a format that ICM can understand. Another global implication is the latency in the data network between two countries. Because ICM relies on heartbeats to verify that all the necessary peripherals are online, it will only wait so long for a heartbeat before it considers the peripheral to be offline. When the latency is too high between the two sites, the peripherals will continuously fight with each other to be the active peripheral. It takes an ICM peripheral about 3 minutes to completely fail over; during this time calls will not be routed.
Since the introduction of Voice Over IP technology, call centers are making a large effort to migrate from TDM environments to Voice over IP networks. VoIP does not add new capabilities to the call center, but instead the technology makes operations cheaper, more cost effective, and less complex. In a VoIP network, a Call Manager replaces the ACD, and a VRU, or Network Voice Response Unit, replaces the traditional VRU. These are peripherals that no longer use phone lines to communicate but instead use the data network. A product called IPCC, or IP Contact Center, takes the place of the traditional ICM. Another trend that is becoming more popular among call centers is the integration of new services that are facilitated by ICM.
These services include Click to chat, Web collaboration, E-mail collaboration, and blended environments using IPCC. Web collaboration allows the call center agent to communicate over the World Wide Web with voice, chat, and data sharing. With this feature an agent can help a customer complete an online document, such as a credit card application, or share the customers Windows desktop with nothing more than a web browser. This type of personalized assistance enhances the customer's experience. Chat sessions are facilitated in the same fashion as a call.
Agents are logged into skill groups that are monitored by ICM. A web server delivers route requests to the Central Controller and ICM checks the availability of the agents before enabling a communication session between the customer and the agent. (Cisco Web Collaboration Option) Also available on many websites today is the ability to send an email to an agent. Similar to click to chat, the agents are moved into a skill specifically designed to only answer customer mail. ICM delivers the email just as it would a call, except that the email is populated on the agent's screen. A blended environment is when an agent handles inbound and outbound calls.
This technology requires the addition of dialer hardware connected to CTI. The Cisco IPCC product implements a true blended environment, capable of accurately predicting call queuing trends which allow agents to re-skill themselves multiple times throughout their shift. ICM has a similar feature, however it could not accurately predict how long an agent would be on an outbound telephone call. As a result the agent was still accounted for in the inbound skill even thought the agent was placing an outbound telephone call to a customer. (Cisco Systems Blended Agent) The world of telecommunications has come a long way over the years and it is evident that there is still a long way to go. There was a time when a phone call required human intervention to establish a connection; now computers are using logic to make intelligent decisions and route the calls for us.
At one time the call center agent had not a clue as to whom he or she was about to speak. Now agents get screens full of data about the customer prior to speaking with them. New ways to communicate with the customer have been invented that save a company's money while giving interaction with their customers the personal touch. All of these added features and more demonstrate just how valuable ICM is to a company's customer service strategy.
Bibliography
Are you running a call centre?" Communicate, February 2004.
Findarticles. com. Retrieved from the Internet 16 July 2005.
Available URL: web m 0 BK / is 2004 Feb / ai n 6149235.
Cisco Blended Agent". Alliance data. com. Available URL: web contact / blended agent. asp". Cisco Systems, Cisco Web Collaboration Option". Cisco. com. Available URL: web Rohde, David. "Geo tel Cashes in on Advanced Call Routing". Network World, June 16, 1997.
Available URL: web qa 3649/is 199706/ai n 8764577.
Stern, Harold P.E. and Samy A. Mahmoud. "Theory of time division multiplexing". Prentice Hall PTR, 2004.