Overview
Answering Machine Detection (AMD) accuracy lies within a competitively sensitive issue, focusing on two aspects: accuracy and latency, and that is the time it takes to identify Determines whether a system call is being answered automatically at the customer end. AMD devices with a Digital Communications Processor (DCP), once properly configured, can have up to 90% ± 5% accuracy. However, it should be noted that the mobile phone and the quality of the phone network may affect this accuracy.
Máy phát hiện Trả lời tự động với Bộ xử lý Truyền thông Kỹ thuật số
The method that Aspect uses to detect the response is called the “beat method”. DCP takes less than 1.5 seconds of network usage to detect an unanswered call. In fact, many times DCP can detect that a person or an answering machine has answered a call within 0.9 seconds of the caller’s first “Hello”.
When DCP determines that the answering machine is not an answering machine, the call will be immediately transferred to the operator to maintain the best service quality. The intelligent AMD algorithm helps to ensure that answering calls is not disconnected while being answered directly by the staff.
Identify competitive advantage
Understanding how the accuracy percentage of AMD is calculated is very important. This accuracy is determined through the calls that have been answered. The formula to determine the accuracy of AMD is:
- The number of the answering machine detected by the system
- Number of automatic answering machines detected by the system + Number of calls arranged for telephone operators
Since the call layout for the enumerator is also used to determine the correctness of AMD, it is important that the call termination codes are correctly identified and that the operator can understand them. There should be two separate call termination codes, one for the operator to use when AMD is active and one when AMD is inactive.
On the contrary, some other competitors use all preset calls, in order to limit the calls that are answered. Since more than 60% of outbound calls are either unanswered, busy or answered, this calculation method will show a significantly higher accuracy rate. If DCP changes this approach, the accuracy AMD will reach 99%+ range.
Another factor to consider is the number of direct, non-answering machine responses. DCP factored this into the correct ratio by looking at the calls arranged for enumerators.
Most systems only allow the user to monitor the phone when switching dials. Aspect’s solutions allow users to manage the DCP’s TV or connection. When monitoring the connection part of the adapter, the user will hear the call being dialed, the phone ring, and also be able to judge the accuracy of the Auto Answer Detector. When monitoring the operator, the user will only hear the calls after they have been classified and forwarded to the operator.
What does a few percentage points really mean?
The following example illustrates the value of the percentage gain of answering machine detection, accomplished through a contact center solution. The table below shows the cost (relative to enumerator lost time) with an answering machine detection rate of 60%.
| Number of calls/ phone number/ hour | 120 |
| Percentage of answering machines/hour | 20% |
| Automatic answering machine | 24 |
| Percentage of answering machines detected | 60% |
| The answering machine is handled by the system | 14.4 |
| The total answering machine is transferred to the telephone operator | 9.6 |
| Total time each answering machine automatically switches to the operator | 25 |
| Total seconds used per hour on answering machines | 240 |
| Telephone number | 75 |
| Cost per hour to pay | $13 |
| Number of hours per day | 12 |
| Number of days per year | 300 |
| Total cost per year | $234,000 |
The following table shows the cost (relative to lost time for enumerators) with an answering machine detection rate of 90%.
| Number of calls/ phone number/ hour | 120 |
| Percentage of answering machines/hour | 20% |
| Automatic answering machine | 24 |
| Percentage of answering machines detected | 90% |
| The answering machine is handled by the system | 21.6 |
| The total answering machine is transferred to the telephone operator | 2.4 |
| Total time each answering machine automatically switches to the operator | 25 |
| Total seconds used per hour on answering machines | 60 |
| Telephone number | 75 |
| Cost per hour to pay | $13 |
| Number of hours per day | 12 |
| Number of days per year | 300 |
| Total cost per year | $58,500 |
The difference in operating the contact center in this example amounts to more than $175,500 per year.
What does a few percentage points really mean? In this example, for each answering machine detection percentage point is increased by $7,020 in the recovered enumerator’s lost productivity.
Conclude
Aspect’s DCP offers the most flexible, intelligent and accurate Auto Answer Detector in the outbound field. The higher the accuracy of answering machine detection, the more significantly the enumerator’s productivity is likely to increase.
Source: Aspect – Compiled by: Lelt
