Telephonic communications – Centralized switching system – Call distribution to operator
Reexamination Certificate
1998-08-17
2002-03-26
Matar, Ahmad (Department: 2642)
Telephonic communications
Centralized switching system
Call distribution to operator
C379S265070, C379S309000
Reexamination Certificate
active
06363145
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates generally to methods and systems for ACD silent call monitoring and, more particularly, the invention relates to a method and system for ACD silent call monitoring utilizing automated voice analysis.
ACD systems are employed in a wide range of customer service environments and provide ACD users with an economical and powerful means for providing customer service. Calls can be automatically routed to ACD agents who possess the skills required for a particular call. Management of a customer service department is facilitated by the ability of a manager or supervisor to monitor calls directed to agents and to analyze performance of the agents. The supervisor, utilizing a single terminal, is able to conveniently switch between different ACD agent calls and thereby directly monitor the job performance of a large number of ACD agents.
An example of an ACD system for providing customer support for computers and peripheral devices might include a greeting which requires the customer to select from among computer related questions regarding printer malfunctions, an inability to run particular applications on the computer, and software installation problems. The option which the customer selects causes the ACD system to transfer the call to a particular agent whose skills match the needs of the customer, as indicated by the selection. Once the call has been answered by an ACD agent, a supervisor can silently monitor the call to evaluate the agent's performance and make recommendations.
If the ACD supervisor becomes familiar with the ACD agents whom the supervisor oversees, the supervisor is able to allocate his time effectively by spending a majority of time on those ACD agents whose skills require more development and spending less time supervising those ACD agents whose skills are already refined. However, if an ACD supervisor is not familiar with the ACD agents he supervises, either because of high agent turnover or because he is new to the job, the supervisor does not have the proper information to focus attention on those ACD agents who most urgently require it. Consequently, the supervisor wastes time monitoring agents who do not require supervision, while agents who urgently require supervision do not receive the attention they require.
A partial solution to the above described problem is provided in U.S. Pat. No. 5,737,405 to Dezonno, which describes an apparatus and method for detecting conversational interruptions in an ACD system, wherein the system includes a detection circuit for detecting when an agent and a customer are concomitantly talking during an incoming call. A caller audio signal detector detects customer audio signals representative of caller speech and an agent audio signal detector detects agent audio signals representative of agent speech. When the caller audio signal detector and the agent audio signal detector detect speech signals simultaneously, an interruption has occurred. Information regarding the interruptions is provided to a supervisor via a supervisor terminal and/or the agent via an agent terminal. The interruption information is presented by showing the agent identification, total call time, agent talk time, and the number of interruptions. Alternatively, the information is presented by showing the averages of total call time, agent talk time, and numbers of interruptions.
Detecting interruptions in incoming ACD calls provides a measure of insight into the performance of an ACD agent. However, there are numerous other indicia which provide a more complete description of agent performance during an incoming call. Furthermore, the Dezonno invention provides for transmitting notification of interruptions to the agent and supervisor terminals as they occur without any processing of the interruption data, or after the call is terminated in the form of a call report which represents the interruption data in a processed format. An ACD supervisor might not be able to effectively utilize the call interruption information in an unprocessed format during the call. Either the supervisor receives unprocessed call interruption data from numerous calls about agent interruptions as they occur, which can easily become overwhelming, or the supervisor receives processed data after the call is terminated, which is obviously too late for the supervisor to utilize during the call.
What is needed is a method and system for providing automated ACD call monitoring which enables a supervisor to utilize information generated by the monitoring during the pendency of the call and which provides a more complete description of agent performance than is currently available.
SUMMARY OF THE INVENTION
A method and system for silent ACD call monitoring utilizing automated voice analysis includes identifying multiple voice data patterns associated with substandard agent performance, monitoring a first call between an ACD agent terminal and a customer terminal to detect the voice patterns, determining whether the number of occurrences of any one of the voice data patterns exceeds an associated predetermined threshold, and notifying the agent terminal and/or a supervisor terminal upon detecting that a threshold number of voice data pattern occurrences has been exceeded.
In a preferred embodiment, the invention is practiced in an ACD system wherein incoming calls from customers are routed to multiple ACD agents. The monitoring of the first call can be performed at any one of multiple sites, including the agent terminal, the supervisor terminal, or another location which has access to both agent voice data and customer voice data.
A processor, such as a digital signal processor (DSP), is utilized to detect the voice data patterns associated with substandard performance. The voice data patterns include a length of silence in conversation between the customer and the agent in excess of a predetermined length which indicates either poor information delivery by the agent or lack of interest on the part of the customer. A conversation volume above a maximum volume level tends to indicate a high frustration level in either the customer or the agent. Changes in voice frequency in excess of a predetermined range during the conversation tend to indicate an emotional exchange. The average length of continuous conversation by the agent also provides information about agent call performance. If the average length of continuous agent conversation is above a maximum threshold, this tends to demonstrate that the agent is talking without paying sufficient attention to the customer. If the average length of continuous agent conversation is below a minimum threshold, this tends to show that the agent is not being responsive to the customer. Interruptions in conversation tend to indicate poor agent performance as well. An interruption of the agent by the customer demonstrates that the customer is frustrated and dissatisfied with the agent's responses, while interruptions by the agent demonstrate that the agent is being impatient and not listening to the customer.
Memory stores voice data pattern thresholds associated with the voice data patterns, such that each voice data pattern has a corresponding threshold. A threshold identifies a maximum number of occurrences of a corresponding voice data pattern which is tolerated during the first call between the agent and the customer. The thresholds can be configured to permit a predetermined number of occurrences for any given time interval during the first call and the number of occurrences permitted by thresholds associated with different voice patterns can differ. For example, the threshold number of agent interruptions per five-minute interval might be greater than the threshold number of occurrences of agent speech having volume above a predetermined level.
The number of occurrences of the voice data patterns are recorded by a counting device, the function of which can be performed by the DSP or a central processing unit (CPU). The number of detected occurrences of the voice data patterns is
Beyda William J.
Gheorghiu Florin M.
Shaffer Shmuel
Skrzynski Mark
Deane, Jr. William J.
Siemens Information and Communication Networks Inc.
LandOfFree
Apparatus and method for automated voice analysis in ACD... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and method for automated voice analysis in ACD..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method for automated voice analysis in ACD... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2872823