JPH06261126A - Operator support type voice interactive device - Google Patents

Abstract

PURPOSE:To provide an operator support type voice interactive device which can use a voice guidance output device for most of conversations including those requiring the judgement of the operator. CONSTITUTION:This device is provided with a storage means 1 to store the voice guidance candidates necessary for conversation and the visual information on each candidate, a visual information display means 2 which shows the visual information to an operator in accordance with the voice guidance candidates stored in the means 1, a visual information selector means 3 which selects the visual information shown by the means 2, and a voice output means 4 which outputs the voice guidance corresponding to the selected visual information in voices.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an assisted voice dialog device, and more particularly, to one corresponding to those uttered by a person skilled in the art who engages in dialogue when the type of uttered contents falls within a certain range. The present invention relates to an operator-assisted voice dialog device in which a voice guidance candidate is prepared in advance, and a speaker selects a voice guidance to be actually output from the candidates to establish a dialogue between an operator and a customer.

[0002]

2. Description of the Related Art Conventionally, in a service agency service by voice conversation within a certain range of utterance content, an operator and a customer directly interact with each other through a voice channel. It is dealt with by repeating utterances. In order to reduce the load on the operator, a method of automating a fixed part of the operator's utterance has been proposed.

FIG. 3 shows an example of a dialogue between an operator and a customer in conventional telephone number guidance. According to the example of FIG.
A method for automating a fixed part of an operator's utterance has been proposed. For example, in the dialogue example shown in FIG.
Nos. 4 and O5 are utterances processed by fixed phrases.
Regarding the utterances of O2 and O3, the operator arbitrarily utters them in order to retrieve the telephone number desired by the customer. Although such an arbitrarily uttered sentence cannot be automated, O6 is an utterance in which the operator conveys the search result to the customer. The utterance of O6 has already been automated even in the currently implemented telephone number guidance service.

Further, according to the method proposed in Japanese Patent Application No. 4-77044, it is possible to automate the utterance portion in which the operator obtains information from the customer by a standard dialogue like O1. O1 is a fixed sentence that is always uttered when a business request is received from a customer, and these are sentences that may be automatically uttered.

[0005]

However, the conversations that can be automated by the above-mentioned conventional method are limited to the fixed conversations that do not need to change the utterance content according to the judgment of the operator, and the utterances of the customer such as O2 and O3. When it is necessary for the operator to judge according to the contents and speak different contents, a voice channel must be provided between the operator and the customer as in the conventional case, and direct dialogue must be made. This causes the following problems.

1. Since automation is limited to fixed-form utterances that do not change according to the operator's judgment, the range of applicable dialogues of the voice output device is narrow and the labor saving effect of the operator is small.

2. If the guidance voice from the voice output device of the automated portion is different from the voice of the operator itself in the other portions, a feeling of discomfort is given to the customer. Alternatively, in order not to give a feeling of strangeness, it takes time and effort to construct another system, such as recording and using the operator's own voice as the guidance voice of the automation part.

3. In order to make an utterance that requires the operator's judgment, it is necessary to provide a voice channel between the operator and the customer, and the operator must perform an operation for that when uttering, which is burdensome.

The present invention has been made in view of the above points,
Resolving the above-mentioned conventional drawbacks, preparing a sentence candidate to be uttered by the operator before the utterance to the operator, and by outputting the guidance selected by the operator to the customer, the utterance that requires the operator's judgment is provided. It is an object of the present invention to provide an operator-assisted voice interaction device capable of enabling the use of a voice guidance output device in almost all operator utterances including.

[0010]

FIG. 1 is a block diagram showing the principle of the present invention.

According to the present invention, in a voice dialogue device that supports a dialogue between a work requester (customer) and a request reception difference operator (operator) in order to receive a work agency service, candidates for voice guidance necessary for the dialogue and each candidate. 1 of the visual information presented by the visual information presenting means 2 and the visual information presenting means 2 presenting the visual information corresponding to the voice guidance candidate of the storage means to the operator. The visual information selecting means 3 for selecting one or a plurality of pieces of information, and the voice output means 4 for outputting the voice guidance corresponding to the visual information selected by the visual information selecting means 3 as a voice.

Further, the visual information presenting means 2 of the operator-assisted voice interaction device of the present invention prioritizes the voice guidance candidates at each utterance time of the operator based on the input information of the operator and the output information of the service device. Priority evaluation means 5 for evaluating, and visual information format changing means 6 for changing the format in which the visual information corresponding to the voice guidance candidate is presented to the operator based on the priority evaluation result of the priority evaluation means.

[0013]

According to the present invention, the operator looks at the guidance displayed on the display, selects the voice guidance from the keyboard according to the utterance content from the customer, and outputs the voice guidance corresponding to the displayed guidance to the customer. Thus, the operator does not have to speak a fixed phrase.

Further, the present invention simplifies the guidance selection by the operator by evaluating the priority of the guidance candidates at the time of the utterance of the operator and displaying them on the basis of the evaluation result.

Further, for the utterance of the operator, the displayed voice guidance is selected, and all the utterances are output from the voice dialogue device for voice output, so that the voice of the operator and the voice from the voice dialogue device are selected. Since there is no confusion with and, the customer does not feel uncomfortable.

[0016]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 2 shows the system configuration of an embodiment of the present invention. In the system shown in the figure, an operator 10 and a customer 20 are connected to a voice dialogue device 40 via a public network 30, and the voice dialogue device 40 is connected to a service device 50.

The voice dialogue device 40 includes a display 41,
The keyboard 42, the visual information storage unit 43, the control unit 44, the voice output unit 45, and the connection control device 46 are included.

The display 41 displays visual information corresponding to the voice guidance candidates to the operator 10.

On the keyboard 42, the operator 10 selects one or more candidates of the visual information of the voice guidance candidates displayed on the display 41.

The visual information storage unit 43 stores candidates for voice guidance necessary for dialogue and visual information corresponding to each guidance candidate on a magnetic disk.

The control unit 44 controls the entire voice dialogue device 40.

The voice guidance reproducing section 45 reproduces and outputs a voice guidance candidate previously recorded in the visual information storage section 43 and designated by the operator. Further, as a device having the same function, a voice synthesizing device which synthesizes and outputs a voice corresponding to a given character string may be used.

The connection control device 46 changes the connection state between the operator 10, the customer 20 and the voice output section 45.

The priority evaluation unit 47 evaluates the priority of each voice guidance candidate at each utterance point of the operator 10 based on the input information of the operator 10 or the output information of the service control device 50. For example, the first utterance requested by the customer 20 is given a priority such that “The number guidance center. Please contact us” is given the highest priority.

The presentation format changing unit 48 uses the display control device used in the existing computer with a display or the like to evaluate the priority of the voice guidance candidates based on the evaluation result of the voice guidance candidates. The presentation format for the operator 10 is changed.

A case where telephone number guidance is provided in the system having the above-described configuration will be specifically described. The visual information storage unit 43 stores the data of typical voice guidance candidates that are frequently spoken by the operator during the dialogue for the telephone number guidance work together with the visual information linked to each guidance candidate in a one-to-one correspondence. I'll do it. As the visual information, it is easy to use a character string indicating the content of the guidance, but any expression form such as a figure can be used as long as the operator 10 can visually identify each guidance.

In a business agent service such as telephone number guidance, the operator's utterance is performed for a specific purpose such as acquisition of information from a customer, and therefore dialogue can be performed with a relatively small amount of voice guidance. It is possible. Table 1 shows examples of guidance candidates for telephone number guidance.

[Table 1]

A series of operations will be described below. When an incoming call is received from the customer 20, the connection control unit 46 connects the output from the customer line 31 to the headset of the operator 10 and the input to the customer line 31 to the voice guidance reproducing unit 45.

Subsequently, the dialogue between the operator 10 and the customer 20 is performed according to the following procedure. (1) The control unit 44 displays a list of visual information corresponding to the voice guidance candidates stored in the visual information storage unit 43 on the display 41. The guidance candidate character strings shown in Table 1 are displayed on the display 41 as they are. (2) The operator 10 selects the most suitable guidance candidate based on the visual information displayed on the display 41 (for example, G1 in Table 1 immediately after an incoming call from a customer). As a selection method, an existing method such as inputting a number associated with a guidance candidate using the keyboard 42 is applied. (3) The control unit 44 sends the voice guidance data of the visual information storage unit 43 to the voice guidance reproducing unit 45 based on the output of the guidance candidate selected by the keyboard 42, and outputs the voice guidance to the customer 20. (4) Since the customer 20 who has heard the guidance speaks to provide customer information, the operator hears it through the headset and inputs the customer information necessary for service to the service device 50. The operator 10 repeats the above-mentioned steps (1) to (4) to perform a dialogue and execute a service for acquiring all necessary customer information.

As described above, the operator 10 simply selects an appropriate voice guidance from the voice guidance candidates displayed on the display 41, and provides the service with almost no direct dialogue with the customer. It becomes possible to do.

In the procedure of (1) above, the display 4
The simplest method for displaying (the visual information indicating) the voice guidance candidate in 1 is to always display the previous candidate. However, in the actual voice dialogue, the possibility that the operator 10 uses each voice guidance at each utterance is not equal, and the progress procedure of the dialogue, the information acquisition situation from the customer 20 of the operator 10, and the service. It changes depending on the execution result.

Therefore, as a second embodiment of the present invention, the priority evaluation unit 47 evaluates the priority of each voice guidance candidate for each utterance time of the operator 10, and changes the presentation format based on the result. In the section 48, the display format of the visual information corresponding to each voice guidance candidate is changed to a format in which only the voice guidance candidates having a higher priority order or a priority order higher than a certain threshold are narrowed down and displayed. By displaying at 41,
It is possible to reduce the effort of the operator 10 selecting the voice guidance.

The method of the second embodiment of the present invention will be described below. It is considered that the probability that each voice guidance is uttered at each utterance time of the operator 10 shows a certain distribution depending on the input state to the service control device 50, the output state from the service control device 50, and the like.

Table 2 shows a table for voice guidance priority evaluation.

[Table 2]

For example, in the conventional dialogue shown in FIG. 3, when the customer information is not acquired immediately after the connection with the customer, the operator speaks with the voice guidance G1 in Table 1 with a probability of 100%. At the time of the utterance O2, the surname and the city / ward name are already acquired in the customer information, and the first name and the town name are not acquired, so it is considered that the voice guidance G2 or G3 is likely to be uttered. The customer information acquisition status of various operators is determined by the service control device 5.
It can be grasped by the state of 0 (the keyboard of the display 41 is used for inputting customer information to the telephone number search device in this embodiment).

Here, the number of times each voice guidance is uttered under each situation is statistically investigated with respect to the information input situation from the keyboard 42 and the situation of the search result, and based on that, each guidance is conducted under each situation. The distribution of the probabilities that U.S. is uttered is stored in the priority evaluation unit 47 in the format shown in Table 1, for example. Then, the input state from the keyboard and the output from the service device at each utterance time of the operator are referred to in this table. It is possible to obtain the priority order of each voice caidance performed.

As described above, the priority of each voice guidance candidate at each utterance point of the operator is evaluated,
By making the presentation format changing unit 48 reflect the presentation method to the operator 10, it is possible to reduce the labor of the operator 10 for selecting the voice guidance.

Further, the method proposed in the present embodiment is applied to Japanese Patent Application No.
In 77044, automation of a standard dialogue as proposed and combination with a direct dialogue channel setting means between an operator and a customer perform a routine dialogue part which does not require operator's judgment by the automation guidance,
It is possible to realize an operator-assisted voice interaction device in which the operator selects and outputs other utterances that require judgment by the operator by the method of the present embodiment, which is more effective in labor saving of the operator.

Furthermore, even if the voice guidance candidates prepared in advance cannot handle the situation, it is possible to deal with the situation by providing a voice channel between the operator and the customer and directly interacting with each other.

[0041]

As described above, according to the present invention, the operator selects the guidance to be output from the presented voice guidance candidates by using the keyboard or the like, and includes the utterance that requires the operator's judgment. Almost all operator utterances can be output using a voice guidance reproducing device to have a dialogue with a customer. As a result, the following effects can be obtained.

1. The operator can carry out the dialogue only by designating the necessary voice guidance at the time of each utterance, and there is almost no need to carry out the direct dialogue with the customer.
Therefore, it is possible to reduce the labor of repeatedly performing the standard dialogue every time the business request is received.

2. It can be used in combination with a part of automation of dialogue by automatic guidance output, and even in the case of utterance that requires an operator's judgment, the dialogue can be performed using the voice output device. Therefore, it is possible to avoid a sense of discomfort experienced by the customer due to the mixture of the guidance voice and the voice of the operator himself / herself during the dialogue.

3. Since the voice output to the customer is always performed via the voice guidance reproducing unit, the operation of setting the voice channel when the operator speaks can be omitted.

[Brief description of drawings]

FIG. 1 is a principle configuration diagram of the present invention.

FIG. 2 is a system configuration diagram of an embodiment of the present invention.

FIG. 3 is a diagram showing an example of interaction between an operator and a customer in conventional telephone number guidance.

[Explanation of symbols]

 1 Storage Means 2 Visual Information Presenting Means 3 Visual Information Selecting Means 4 Voice Output Means 5 Priority Evaluating Means 6 Visual Information Format Changing Means 10 Operators 20 Customers 30 Public Networks 40 Voice Conversations 41 Display 42 Keyboard 43 Visual Information Storage 44 Control Part 45 Voice guidance reproducing part 46 Connection control part 47 Priority evaluation part 48 Presentation format changing part 50 Service control device

Continuation of front page (51) Int.Cl. ⁵ Identification code Office reference number FI Technical display location G10L 3/00 571 H 9379-5H

Claims (2)

[Claims] 1. A voice dialogue device for supporting a dialogue between a work client (customer) and a request recipient (operator) for accepting a business agent service, supporting voice guidance candidates and each candidate necessary for the dialogue. A storage unit for storing the visual information; a visual information presentation unit for presenting the visual information corresponding to the voice guidance candidate of the storage unit to the operator; and one of the visual information presented by the visual information presentation unit. An operator-assisted voice, comprising: visual information selection means for selecting one or more pieces of information; and voice output means for outputting, as voice, voice guidance corresponding to the visual information selected by the visual information selection means. Dialogue device. 2. The visual information presenting means, based on the input information of the operator and the output information of the service device, a priority order evaluating means for evaluating the priority order of each of the voice guidance candidates at each utterance time point of the operator, Based on the priority evaluation result of the priority evaluation means,
The operator-aided voice interaction apparatus according to claim 1, further comprising visual information format changing means for changing a format of visual information corresponding to the voice guidance candidate presented to an operator.