CN114841128B

CN114841128B - Business interaction method, device, equipment, medium and product based on artificial intelligence

Info

Publication number: CN114841128B
Application number: CN202210335800.3A
Authority: CN
Inventors: 房莹; 孙孟尧
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2022-03-31
Filing date: 2022-03-31
Publication date: 2023-06-20
Anticipated expiration: 2042-03-31
Also published as: CN114841128A

Abstract

The disclosure provides a business interaction method, device, equipment, medium and product based on artificial intelligence, relates to the technical field of data processing, and particularly relates to the field of Natural Language Processing (NLP). The specific implementation scheme is as follows: if a form filling instruction is detected, determining the input work form text and a field to be filled in the form; the work order text is analyzed by a calling model, and fields included in the work order text and attribute values corresponding to the fields are obtained; matching the field with the field to be filled, and filling an attribute value corresponding to the field matched with the field to be filled into an input frame corresponding to the field to be filled; displaying the completed form. After the form filling instruction is detected, the work order text is analyzed through the model, and the fields in the work order text and the attribute values corresponding to the fields are obtained, so that automatic filling of the form is completed, and the filling efficiency of the form is improved.

Description

Business interaction method, device, equipment, medium and product based on artificial intelligence

Technical Field

The present disclosure relates to the field of artificial intelligence, and more particularly to natural language processing, which may be applied in government service, intelligence, and smart city scenarios.

Background

Customer service personnel in the hotline industry are responsible for answering calls and filling out forms according to requirements so as to record and reflect event contents. And then the dispatcher reviews and dispatches the filled form to drive the subsequent flow. Filling out each field in the form item by item requires customer service personnel to have firm business knowledge. However, when the traffic volume increases, it is difficult to ensure the quality and efficiency of form filling at the same time.

Disclosure of Invention

The disclosure provides a business interaction method, device, equipment, medium and product based on artificial intelligence.

According to an aspect of the present disclosure, there is provided an artificial intelligence based service interaction method, including: if a form filling instruction is detected, determining the input work form text and a field to be filled in the form; the work order text is analyzed by a calling model, and fields included in the work order text and attribute values corresponding to the fields are obtained; matching the field with the field to be filled, and filling an attribute value corresponding to the field matched with the field to be filled into an input frame corresponding to the field to be filled; displaying the completed form.

According to another aspect of the present disclosure, there is provided an artificial intelligence based service interaction apparatus, including: the system comprises a form filling instruction detection unit, a form filling instruction detection unit and a form filling unit, wherein the form filling instruction detection unit is used for detecting a form filling instruction and determining a work order text and a field to be filled which are input in a form; the analysis unit is used for calling a model to analyze the work order text to obtain fields contained in the work order text and attribute values corresponding to the fields; the matching unit is used for matching the field with the field to be filled in, and filling the attribute value corresponding to the field matched with the field to be filled in the input frame corresponding to the field to be filled in; and the display unit is used for displaying the filled-in form.

According to still another aspect of the present disclosure, there is provided an electronic apparatus including: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method.

According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the method.

According to another aspect of the present disclosure, there is provided a computer program product comprising a computer program which, when executed by a processor, implements the described method.

It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.

Drawings

The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

FIG. 1 is a flow chart of an artificial intelligence based business interaction method provided in accordance with an embodiment of the present disclosure;

FIG. 2 is a schematic diagram of a form to be filled in provided in accordance with an embodiment of the present disclosure;

FIG. 3 is a schematic diagram of a training worksheet element extraction model provided in accordance with an embodiment of the present disclosure;

FIG. 4 is a flow diagram of identifying an incident address using a worksheet element extraction model provided in accordance with an embodiment of the present disclosure;

FIG. 5 is a schematic diagram of an address normalization process provided in accordance with embodiments of the present disclosure;

FIG. 6 is a flow chart of a method of modifying a form provided in accordance with an embodiment of the present disclosure;

FIG. 7 is a schematic diagram of a homogeneous form search service provided in accordance with an embodiment of the present disclosure;

FIG. 8 is a schematic diagram of a co-object query service provided in accordance with an embodiment of the present disclosure;

FIG. 9 is a schematic diagram comparing the present disclosure with related art provided in accordance with an embodiment of the present disclosure;

FIG. 10 is a schematic flow chart of a pick-and-turn form provided in accordance with an embodiment of the present disclosure;

FIG. 11 is a schematic diagram showing a rejection proposal provided in accordance with an embodiment of the present disclosure;

FIG. 12 is a block diagram of an artificial intelligence based business interaction device, according to an example embodiment;

FIG. 13 is a block diagram of an electronic device for implementing an artificial intelligence based business interaction method of an embodiment of the present disclosure

Detailed Description

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

The method and the device can be applied to the scene of filling the form, and particularly applied to the scene of filling the form according to the call content after a customer service person (telephone operator) receives the call. The present disclosure may also be applied to a scenario in which a form is filled in according to content reflected by a voice description and/or content reflected by a text description by a user is received in a specific scenario. It should be noted that the call content may be content recorded by a customer service person, or may be text information converted from voice information by monitoring voice call. For convenience of description, the following embodiments will be described by taking a scenario in which a customer service person answers and records call contents in real time, and fills out a form according to the recorded call contents.

After receiving the call, the customer service personnel in the hotline industry not only needs to solve the citizen problem according to a certain telephone skill and tolerance, but also needs to sort, record and fill in the form (the form content comprises the fields of an event title, an event occurrence place, an event occurrence time, an event related subject, an event classification, an event affiliated department, an event keyword and the like). The form acts as a carrier for the phone to reflect the problem throughout the process. And from the time of answering a call by a customer service person to the time of completing the item-by-item filling of the whole form content, the customer service person needs to rely on solid business knowledge (such as a matter classification list, a dispatch manual and the like). For the customer service personnel, the entire process of filling out the form is time consuming and labor intensive, typically requiring 3-5 minutes. Particularly, when the telephone traffic is increased, it is difficult to ensure the quality and efficiency of form filling at the same time. If the form is filled out with low efficiency and/or poor quality, the effects of the follow-up dispatch, the underwriting treatment and the business analysis links of the form will be directly affected.

In the related art, aiming at a filling form scene of the government service hotline industry, a menu item or an input box built in a system is provided, and a user for filling the form manually inputs or selects corresponding contents from a drop-down menu so as to finish item-by-item filling of various contents of the form. However, for users who fill out forms, a large number of knowledge books still need to be memorized and understood, which is inefficient and error-prone. When the telephone traffic is increased, new customer service personnel are introduced, and the time cost and the money cost of on-duty training are increased. On the other hand, after the form is filled, the form filling content is required to be audited by a form dispatcher, and the form is signed and transferred based on the auditing result, namely, the form is forwarded to the affiliated processing department or the returned form. The related art has two processing methods for signing and transferring the form, in which the first processing method requires that the signing and transferring reason (refusing signing reason or confirming signing and transferring) be manually input. The method relies on the understanding of the business knowledge by the order taker, but the content filled in after the understanding by the manual memory is easy to be wrong, and for the same problem, there is an objective phenomenon that different people understand the deviation to mislead the handling of the upstream business. The second processing mode directly carries out automatic signing and transferring according to the preset dispatching rules. The second processing mode reduces manpower energy, and can realize automatic signing and transferring of the form through a preconfigured rule, but the process defaults to direct dispatching, and has the defects of lacking a manual checking step and being incapable of supporting service denial.

In view of this, the present disclosure provides a business interaction method based on artificial intelligence, which is based on event normalization technology, text classification technology and sequence labeling technology in NLP natural language processing, and analyzes recorded call content when a form filling instruction is detected. Form filling is completed based on the analysis result. According to the method and the system, the customer service personnel can be assisted to complete filling and signing of each field in the form rapidly under the conditions of no introduction of new customer service personnel and no long-time training. The present disclosure advances the handling and resolution of citizen incoming calls in a low cost, portable replication approach.

In the following embodiments of the present disclosure, a customer service person who fills out a form and a dispatcher who reviews the form are collectively referred to as a user.

The following embodiments of the present disclosure will illustrate methods of processing forms provided by the present disclosure with reference to the accompanying drawings.

FIG. 1 is a flow chart of an artificial intelligence based business interaction method provided in accordance with an embodiment of the present disclosure; as shown in fig. 1, the service interaction method based on artificial intelligence provided by the present disclosure includes the following steps S101 to S104.

In step S101, if a form filling instruction is detected, the inputted work form text and the field to be filled in the form are determined.

In the present disclosure, function options may be set on a page filled in a form, and a mapping relationship between the function options and the filling in of the form may be set. If the user clicking the function option is detected, determining that a form filling instruction is detected.

In step S102, the model is invoked to parse the work order text, so as to obtain the fields included in the work order text and the attribute values corresponding to the fields.

The present disclosure calls a model to parse the work order text. The present disclosure pre-trains multiple models for parsing work order text according to fields to be filled in a form. Each model outputs one or more fields and attribute values corresponding to the fields. And inputting the work order text into a plurality of models, analyzing the work order text by each model, and outputting the corresponding fields and attribute values.

In step S103, the fields are matched with the fields to be filled in, and the attribute values corresponding to the fields matched with the fields to be filled in are filled in the input boxes corresponding to the fields to be filled in.

And matching the fields output by the model with the fields to be filled in the form to be filled in, and completing filling in the form.

In step S104, the completed form is displayed.

After the form filling instruction is detected, the work order text is analyzed through the model, and the fields in the work order text and the attribute values corresponding to the fields are obtained, so that automatic filling of the form is completed, and the filling efficiency of the form is improved.

According to the method and the system, the fields and the attribute values corresponding to the fields are analyzed from the work sheet text through the model and are used for automatically filling the form, so that the skill requirements on business personnel filling the form can be reduced, the personnel cost is reduced, and the form filling quality and efficiency are improved. The method for analyzing the work order text by calling the model in the disclosure to obtain the fields and the attribute values corresponding to the fields, which comprise the following steps: and calling a target model corresponding to the field to be filled based on the field to be filled, and inputting the work order text into the target model to obtain the field and the attribute value corresponding to the field included in the work order text. The object model in the disclosure comprises a first object model and a second object model, wherein the input of the first object model is work order text, and the output of the first object model is a field and an attribute value corresponding to the field. The input of the second target model is the output of the first target model, and the output is a field and an attribute value corresponding to the field.

In one embodiment, taking the form to be filled in of the citizenship line shown in fig. 2 as an example, the fields to be filled in the form to be filled in include a prosecution title, a prosecution subject, a prosecution address, a prosecution type, a matter classification, a hotword tag, and a affiliated department. The content in the citizen appeal is the inputted work order text. And inputting the work order text into a work order element extraction model to analyze, and outputting the attribute value of the title of the requirement, the attribute value of the related subject and the attribute value of the incident address. And inputting the worksheet text into a worksheet classification model for analysis, and outputting the attribute value of the item classification. And analyzing the attribute value input address normalization model corresponding to the event address, and outputting the attribute value corresponding to the normalized event address. And inputting the work order text into a hotword label matching model for analysis, and outputting a hotword label and an attribute value corresponding to the hotword label. And inputting the event address into the department identification model to analyze, and obtaining the attribute value corresponding to the department.

And the target model is utilized to analyze the work order text, so that the fields and the attribute values corresponding to the fields contained in the work order text can be rapidly determined, and the efficient completion of the form filling is facilitated.

In the embodiment of the disclosure, in order to implement the identification of the address of the incident by the work order element extraction model, the work order element extraction model is trained by using the process shown in fig. 3. As shown in fig. 3, the work order element extraction model is trained using annotation data obtained by manual annotation and machine-assisted annotation. The present disclosure is annotated with the grammatical knowledge of whether address information in the text is a "place-and-phrase" or a non-place-and-phrase. Through a large number of data labels, a label dataset is generated. And training the relic model offline by using the labeling data set, and performing model evaluation. The training text model is adopted to realize the task of element extraction-address identification. The trained text center model can automatically analyze the context information of the input text and output the attribute value which is most in line with the field incident address.

In an embodiment of the present disclosure, outputting an attribute value of an event address using a training-completed work order element extraction model includes: and identifying the work order text, if the address information is identified in the work order text, determining grammar information between the address information and the middle event described by the work order text, otherwise, outputting the event address, and setting the attribute value of the event address as a null value. If the grammar information between the address information and the events in the work order text meets the preset grammar rule, determining that the event address exists, otherwise, outputting the event address, and if the attribute value of the event address is null. If there are two or more address information, the address information of the first event is used as the target event address and the attribute value used as the event address is output. If there is one address information, the address information is used as the attribute value of the address.

In one example, a process of identifying an issue address by the work order element extraction model will be described with reference to fig. 4.

As shown in fig. 4, a claim worksheet text is input, and address information of the input text segment is identified. If no address information exists, the output address is "null". If the address information exists, judging whether the address information is used as an address word of the event. In this example, the judgment rules are classified into a complaint report type complaint work order-incident address labeling rule and a consultation recourse type litigation work order-incident address labeling rule. The complaint report type complaint work order-incident address labeling rule comprises that if address information in a work order text meets the sentence patterns of a preset main predicate structure, a main system table structure and a label structure, the address information is an address word. If the address information is not an address word, the output address is "null". If the address information is an address word and there are a plurality of address information, the address information of the first event is outputted as an attribute value of the address of the event. If the address information is an address word and there is unique address information, the unique address information is used as an attribute value of the address in question.

For example, the work order text is "citizen consultation, epidemic prevention and control policy from the bead sea to the Beijing, and nucleic acid detection point positions near the hundred degree scientific and technical garden in the Beijing lake area". The work order text is analyzed by using the calling model of the disclosure, and the analysis result is as follows:

Work order classification: medical health-disease control epidemic prevention (New crown epidemic prevention and control)

The unit of undertaking: horse-tie street; ganjiakou street

A related subject: - -

The address of the event: hundred degree scientific and technological garden in Beijing sea lake area

Time of occurrence: - -

Event topic: epidemic prevention and control policy from the bead sea to the Beijing, and the positions of nucleic acid detection points near the Beijing beaten hundred degrees scientific and technical garden.

In order to facilitate determination of the department in this disclosure, attribute values of the event address are normalized. The mapping of address completion and administrative areas is realized, and the mapping is used for the subsequent service drop point falling map or dispatch. For example, "around a mikan subway station" should be classified into "around a beijing city-sea lake area-mikan street-mikan subway station". The related address is 'Beijing city-sea lake area-Mallotus street-Mallotus subway station vicinity'.

In the embodiment of the present disclosure, a procedure of performing address normalization processing is described with reference to fig. 5. As shown in fig. 4, the input "event address" refers to an attribute value of a field event address. And calling an address normalization model, inputting the attribute value of the event address into the normalization model, and outputting the normalization address. In the present disclosure, the attribute value of the incident address is a normalized address obtained by normalizing the incident address output by the work order element extraction model.

When the identification is carried out in the normalization model, a plurality of normalization addresses meeting the conditions are found in the address library, and then the normalization address with the highest confidence is determined according to the context understanding. For example, address "Li Gucun" is available in many places such as Shaanxi, western Ann, chongqing, etc., and is determined in conjunction with the address information mentioned in the context. If the context repeatedly refers to western or western street, then Li Gucun is determined to be western, i.e. return address is shanxi province-western city-bridge area-new street-Li Gucun.

The forms automatically filled in by the present disclosure may not conform to the user's expectations, and the present disclosure provides a way of modification. That is, the user can quickly fill the selected character string in the work order text into the input box of the corresponding field. Thus, the automatically filled form can be made to conform to the user's expectations.

The present disclosure illustrates, via fig. 6, a method of modifying a form provided by the present disclosure. FIG. 6 is a flow chart of a method of modifying a form provided in accordance with an embodiment of the present disclosure; as shown in fig. 6, the method for modifying a form provided by the present disclosure includes the following steps S601-S603.

In step S601, a selected character string in the work order text is detected.

In step S602, a field corresponding to the character string is determined, and a first prompt message including the field is displayed on the character string.

And obtaining a character string input model and determining a field corresponding to the character string. The first hint information in this disclosure includes "click add XXX".

In step S603, clicking the first prompt message is detected, and the character string is filled into the input box corresponding to the field.

And if the clicking of the first prompt information is detected, filling the character strings into the input boxes corresponding to the fields according to the mapping relation. The work sheet text comprises a plurality of addresses, the address information which is expected to be filled in the corresponding input box of the address is selected by a user, wherein the address information is wrong in the address of the form after automatic filling. And if the model identifies that the selected character string is the address information, displaying first prompt information (event address) on the selected character string.

The method and the device have the advantages that the first prompt information is displayed on the character string selected by the work order text, and the selected character string is filled into the corresponding input box after the first prompt information is clicked. Thus, the shortcut mode supports the user to modify the automatically filled form, and the flexibility of form filling is increased.

As shown in fig. 2, assuming that the user selects the text string of "north-free road xx village xx south-road side of Beijing city by the sea lake area Su Gu lump," the system automatically recognizes that the selected text is a "address of issue" label, and prompts the "address of issue" in the upper right corner, the first prompting information can be clicked by the user, after the user clicks the first prompting information, a string of text content selected by the user can be filled into the corresponding text input box, and the result of automatic filling of the previous machine can be covered. After detecting the deselected continuous text segment, the first prompt message also disappears.

Based on any embodiment, the disclosure further provides auxiliary services such as similar form search (class recommendation), same-number electrician form display, related policy query and the like in order to assist users to fill forms better or provide high-quality services for consultants and/or complaints. The following embodiments will explain an auxiliary service provided by the present disclosure with reference to the accompanying drawings.

The similar form search service provided by the present disclosure includes: detecting a search instruction, and determining input search content, wherein the input search content comprises keywords of a work order text; inputting the input search content into a recommendation model to obtain a first form with similarity meeting requirements with the worksheet text; determining a search result display rule; and displaying the obtained first form according to the search result display rule.

As shown in fig. 7, in one example, the left side of the page is the fill area and the right side is the auxiliary area. It should be noted that the page display in fig. 7 is only an example, and the present disclosure is not limited to the page layout format. The present disclosure provides a homogeneous form search service in an auxiliary area, in which a search box is provided on a page, in which a user can input keywords. To facilitate guiding the user, "enter case keywords," multiple keywords are separated by spaces, may be prompted in a search box. The system will automatically call the interface (Application Programming Interface, API) of the recommendation model to find the form in the form library that matches the search content based on the search content entered by the user. Recommending cases similar to the currently input search content to the user, and assisting the user in completing the filling of the work order content. The present disclosure may employ forms that preferentially display the same number when displaying search results.

Through the search service provided by the disclosure, a convenient search function is provided for a user. The method can provide references for users when answering consultation questions or query progress, and provide reply efficiency for the users.

In this disclosure, a same-object query service is also provided, taking the city long hotline as an example, and the same-object query service mainly aims at single-job single query with the same number. The co-object query service in this disclosure includes: if the same object query instruction is detected, the input object to be queried is determined. The object to be queried in the present disclosure includes a phone number in the work order text and a subject of the work order text. Determining a second form with an association relation with the object to be queried, and clustering the second form according to a preset clustering rule; and displaying the clustering result according to the display rule. As shown in FIG. 8, a user can obtain a unified case aggregation result integrated by the system for the user through the same-object query service, and each work order can be ordered according to a time sequence for facilitating the reading of the user, so that the user is assisted to quickly know the historical appeal condition of the object, and the business handling is completed efficiently.

In one embodiment, the aggregation services provided by the present disclosure employ a work order classification model, a work order element extraction model, and an event normalization technique. The present disclosure provides an aggregation service to aggregate together different representations of citizen power orders. For example, the ms and mr. Who make phone calls respectively complain about the noise disturbance problem of a certain business, but the text expression is different. When the incoming call worksheets of the plum women and the Mr. are polymerized, the worksheets are filtered by using the worksheet classification model, and the filtered worksheets are polymerized, so that the polymerization effect can be improved.

According to the method and the system for inquiring the objects, through the same-object inquiring service, when the same-object inquiring instruction is detected, the forms associated with the objects to be inquired are determined, the forms are clustered, and clustering results are displayed to the user, so that the service handling efficiency of the user is improved.

After the user clicks the "submit form" button in the present disclosure, the system automatically saves the form content and distributes it to the signature handling personnel for the next circulation of the work order.

The business interaction process based on artificial intelligence in the embodiment of the disclosure comprises two parts of form filling and form assignment. The method and the device can complete automatic filling of the form according to the input work order text and the form filling instruction. After the form is filled out, the intelligent form dispatching method and the intelligent form dispatching system can realize intelligent form dispatching. And calling a pre-trained model to audit the form from multiple dimensions, determining the content which exists in the form and does not accord with the assignment rule, and generating a rejection suggestion for the user to select. And receiving an instruction of clicking the refusal sign suggestion by the user, automatically filling the refusal sign establishment into an refusal sign reason input box, and improving the form auditing efficiency without editing the refusal sign reason.

In order to facilitate understanding of the form processing method provided by the present disclosure, the embodiment of the present disclosure uses fig. 8 as an example to illustrate the difference between the present disclosure and the related art. FIG. 9 is a schematic diagram comparing the present disclosure with related art provided in accordance with an embodiment of the present disclosure; as shown in fig. 9, the left side is a related art business interaction flow based on artificial intelligence, and the right side is form processing by the present disclosure through intelligent interaction of front end and back end. The business interaction flow based on artificial intelligence in the related art comprises logging in an office system, entering a work order registration page, filling in various contents of a form, and then submitting the form. And the system distributes the form to form signing and transferring service personnel, and the service personnel conduct signing and transferring judgment according to the service rules and knowledge to determine whether the form filling information is correct. If the form filling information is correct, the signing and transferring is confirmed, and the system executes form dispatching and transferring. If the form filling information is wrong, filling out a refusal reason, and returning the form after the manual examination is passed. Compared with the related art, the method adds the [ one-key filling ] option in the [ work order registration ] page, and clicks the [ one-key filling ] option after a user inputs the work order text. The system detects that the [ one-touch pad ] option is triggered, invoking the backend model. And inputting the worksheet text into the model, and outputting a research and judgment result. And filling attribute values corresponding to the fields in the form into input boxes corresponding to the fields in the form based on the attribute values corresponding to the fields in the research and judgment result. In the form signing and transferring scene, a business person does not need to sign and transfer judgment according to business rules and knowledge, but calls a model to review the form to be reviewed, and the signing refusing reason is output and displayed. And if the user clicks to confirm the refusal reason, filling the refusal reason confirmed by the user into the corresponding input box, and returning the system execution form after the manual verification is passed. The models in the disclosure include a work order element extraction model, a work order classification model, an address normalization model, a hotword tag matching model, a belonging part recognition model, and the like. The above model is encapsulated and APIs are provided in the present disclosure to facilitate calls.

As can be seen from fig. 9, the present disclosure completes form filling by calling the model after receiving an operation instruction to fill in the form, and completes the examination of the form by the model, and gives out a rejection suggestion. The method has the advantages that the user does not need to input the form content, the reason of refusal sign editing is not needed, the professional skill requirement on the user is reduced, the quality and efficiency of form filling are improved, and the scene of telephone traffic surge can be more gracefully dealt with. Compared with the automatic signing and transferring of the form realized through a preconfigured rule in the related art, the method and the device generate the alternative signing refusing suggestion, complete signing and transferring after manual auditing, and overcome the defect that service signing refusing cannot be supported in the related art.

The upper dashed box in fig. 9 is the intelligent form filling process, and the lower dashed box is the intelligent assignment handling process. It should be noted that, the intelligent form filling process and the intelligent dispatch handling process provided in the present disclosure may be used independently. By way of example, the method and the system can effectively complete quick response of citizen consultation help-seeking problems, quick filling and signature-transfer handling of complaint report suggestion forms for users in the government-service hotline industry. Through the interactive tool of intelligent form filling and intelligent dispatch, the elements of the input work form text can be analyzed and normalized by calling an NLP algorithm model, category labels are researched and judged, and one-key quick filling of various contents of the form is realized. According to the method and the system, additional services are added on the basis of the scheme, so that active searching and recommending of the same or similar forms are achieved, users can directly answer citizen problems and fill out form content references, and flexible signing and conversion and refusal signing reason quick filling of the forms are facilitated.

The disclosed embodiment uses fig. 10 as an illustration of a process for signing a form. FIG. 10 is a schematic flow chart of a pick-and-turn form provided in accordance with an embodiment of the present disclosure; as shown in fig. 10, the process of signing up the form includes steps S1001-S1004.

In step S1001, if it is detected that the tab page is entered, a form to be audited is determined, and the form to be audited is a form completed based on the worksheet text.

In step S1002, the call model performs synchronous auditing on the form to be audited from multiple angles, and determines a character string which does not conform to a preset rule and a rejection suggestion corresponding to the character string in the form to be audited.

The method and the device are used for learning the rules of the signing and transferring of the related forms in advance by the models applied to the signing and transferring scene, synchronously auditing the forms to be audited from multiple dimensions through the models, and enabling the auditing angle to be more comprehensive. The service prompt is carried out in a friendly interaction mode of the refusing list suggestion, and the method can assist in manually and rapidly completing the confirmation of the signing of the work list or the order or refusing the list.

In step S1003, a character string is marked in the form to be checked, and a second prompt message containing a sign rejection suggestion corresponding to the character string is displayed in a preset display area.

As shown in fig. 11, the details of the work order are marked with character strings which do not conform to the rule, and the second prompt information is displayed in the refusal to sign suggestion area. The second prompt information is connected with the marked character strings through line segments. And reflecting the mapping relation between the second prompt information and the labeling character string through the connecting line. The second prompt message is provided with a [ refusal sign ] option. And if the user clicks the item [ refusing the signature ] is detected, filling the refusing suggestion corresponding to the refusing item into an input box. The steps of tedious judgment and refusal reason filling of the user are omitted, and the working efficiency can be greatly improved.

In step S1004, an instruction for confirming the second prompt information is detected, and the rejection suggestion corresponding to the second prompt information is filled into the rejection reason input box.

If the check-in instruction is detected, determining the affiliated department in the form to be checked, and forwarding the form to be checked to the affiliated department. If the return instruction is detected, the form to be checked with the refusal sign reason is returned. As shown in fig. 11, the user clicks the "back" and "confirm assignment" buttons, and corresponding prompts are popped up above the page, and at the same time, the user automatically switches to the next work order. If the user clicks the "back" button, the page popup prompts: "returned". If the user clicks the "confirm assignment" button, the page popup prompts: "has been assigned to [ belonging to ] the department that is the final filling result of the user. The signing and handling link is used as a key link in the work order flow and the business efficiency and quality of the link are improved, and the overall business efficiency is also greatly improved.

According to the method and the device, the form to be audited is audited in multiple dimensions through the call model, the auditing efficiency can be improved, when the second prompt information containing the refusal advice corresponding to the character string is clicked, the refusal advice filling field is an input frame corresponding to the refusal reason, the time for editing the refusal reason by a user is shortened, and the form refusal reason conversion efficiency is further improved.

The interaction logic provided by the present disclosure can be multiplexed in any business scenario requiring to sign-up and dispatch on any business process, such as a city long hot line work order sign-up and order sign-up link. The present disclosure only takes the urban hot line scenario as an example, and the intelligent interactive service provided by the present disclosure is not limited thereto.

Based on the same conception, the embodiment of the disclosure also provides a business interaction device based on artificial intelligence.

It may be appreciated that, in order to implement the above-mentioned functions, the service interaction device based on artificial intelligence provided in the embodiments of the present disclosure includes a hardware structure and/or a software module that perform respective functions. The disclosed embodiments may be implemented in hardware or a combination of hardware and computer software, in combination with the various example elements and algorithm steps disclosed in the embodiments of the disclosure. Whether a function is implemented as hardware or computer software driven hardware depends upon the particular application and design constraints imposed on the solution. Those skilled in the art may implement the described functionality using different approaches for each particular application, but such implementation is not to be considered as beyond the scope of the embodiments of the present disclosure.

FIG. 12 is a block diagram illustrating an artificial intelligence based business interaction device, according to an exemplary embodiment. Referring to fig. 12, the apparatus 1200 includes a detection fill instruction unit 1201, an parsing unit 1202, a matching unit 1203, and a display unit 1204.

A test filling instruction unit 1201, configured to determine a work form text and a field to be filled that are already input in a form if a form filling instruction is detected; the parsing unit 1202 is configured to invoke a model to parse the work order text, so as to obtain a field included in the work order text and an attribute value corresponding to the field; the matching unit 1203 is configured to match a field with a field to be filled, and fill an attribute value corresponding to the field matched with the field to be filled into an input box corresponding to the field to be filled; and a display unit 1204 for displaying the completed form.

In one implementation manner, the parsing unit 1202 is configured to call, based on a field to be filled, a target model corresponding to the field to be filled, where the target model includes a first target model and a second target model, the input of the first target model is a work order text, the output is a field and an attribute value corresponding to the field, the input of the second target model is the output of the first target model, and the output is a field and an attribute value corresponding to the field; and inputting the work order text into the target model to obtain the fields and the attribute values corresponding to the fields included in the work order text.

In one embodiment, the apparatus 1200 further includes a modifying unit 1205 for detecting a selected character string in the work order text; determining a field corresponding to the character string, and displaying first prompt information containing the field on the character string; and detecting clicking the first prompt information, and filling the character strings into the input boxes corresponding to the fields.

In one embodiment, the apparatus 1200 further includes a search unit 1206 for detecting a search instruction, determining that search content has been input, the input search content including keywords of a work order text; inputting the input search content into a recommendation model to obtain a first form with similarity meeting requirements with the worksheet text; determining a search result display rule; and displaying the obtained first form according to the search result display rule.

In one embodiment, the apparatus 1200 further includes an on-object query unit 1207, configured to determine, if the on-object query instruction is detected, an input object to be queried, where the object to be queried includes a phone number in a work order text and a theme of the work order text; determining a second form with an association relation with the object to be queried, and clustering the second form according to a preset clustering rule; and displaying the clustering result according to the display rule.

In one embodiment, the apparatus 1200 further includes a signing unit 1208 configured to determine, if it is detected that the signing page is entered, a form to be audited, where the form to be audited is a form completed based on the worksheet text; synchronously auditing the form to be audited from multiple angles by calling a model, and determining character strings which do not accord with preset rules and refusal sign suggestions corresponding to the character strings in the form to be audited; marking character strings in the form to be checked, and displaying second prompt information containing refusal advice corresponding to the character strings in a preset display area; and detecting an instruction for confirming the second prompt information, and filling the refusal sign suggestion corresponding to the second prompt information into the refusal sign reason input box.

The specific manner in which the various modules perform the operations in the apparatus of the above embodiments have been described in detail in connection with the embodiments of the method, and will not be described in detail herein.

In the technical scheme of the disclosure, the acquisition, storage, application and the like of the related user personal information all conform to the regulations of related laws and regulations, and the public sequence is not violated.

According to embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

Fig. 13 illustrates a schematic block diagram of an example electronic device 1300 that can be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 13, the apparatus 1300 includes a computing unit 1301 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM) 1302 or a computer program loaded from a storage unit 1308 into a Random Access Memory (RAM) 1303. In the RAM 1303, various programs and data required for the operation of the device 1300 can also be stored. The computing unit 1301, the ROM 1302, and the RAM 1303 are connected to each other through a bus 1304. An input/output (I/O) interface 1305 is also connected to bus 1304.

Various components in device 1300 are connected to I/O interface 1305, including: an input unit 1306 such as a keyboard, a mouse, or the like; an output unit 1307 such as various types of displays, speakers, and the like; storage unit 1308, such as a magnetic disk, optical disk, etc.; and a communication unit 1309 such as a network card, a modem, a wireless communication transceiver, or the like. The communication unit 1309 allows the device 1300 to exchange information/data with other devices through a computer network such as the internet and/or various telecommunication networks.

The computing unit 1301 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 1301 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1301 performs the various methods and processes described above, such as the artificial intelligence based business interaction method. For example, in some embodiments, the artificial intelligence based business interaction method may be implemented as a computer software program tangibly embodied on a machine-readable medium, e.g., storage unit 1308. In some embodiments, part or all of the computer program may be loaded and/or installed onto the device 1300 via the ROM 1302 and/or the communication unit 1309. When the computer program is loaded into RAM 1303 and executed by computing unit 1301, one or more steps of the artificial intelligence based business interaction method described above may be performed. Alternatively, in other embodiments, computing unit 1301 may be configured to perform the artificial intelligence based business interaction method in any other suitable manner (e.g., by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), and the internet.

The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server may be a cloud server, a server of a distributed system, or a server incorporating a blockchain.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel or sequentially or in a different order, provided that the desired results of the technical solutions of the present disclosure are achieved, and are not limited herein.

The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.

Claims

1. An artificial intelligence based business interaction method comprises the following steps:

if a form filling instruction is detected, determining the input work form text and a field to be filled in the form;

the work order text is analyzed by a calling model, and fields included in the work order text and attribute values corresponding to the fields are obtained;

matching the field with the field to be filled, and filling an attribute value corresponding to the field matched with the field to be filled into an input frame corresponding to the field to be filled;

Displaying the filled form;

the field comprises an event address, the model comprises an element extraction model and an address normalization model, the element extraction model is used for analyzing and obtaining the address attribute of the work order text, and the address normalization model is used for obtaining an attribute value corresponding to the event address after normalization of the address attribute;

and (3) analyzing the work order text by calling a model in the following way to obtain an attribute value corresponding to the event address:

and calling the model to identify the work order text, and taking the address information as an attribute value of the incident address if the address information is identified in the work order text and grammar information between the address information and events in the work order text meets a preset grammar rule.

2. The method of claim 1, wherein the parsing the work order text by the calling model to obtain the fields included in the work order text and the attribute values corresponding to the fields includes:

invoking a target model corresponding to the field to be filled based on the field to be filled, wherein the target model comprises a first target model and a second target model, the input of the first target model is the work order text, the output is a field and an attribute value corresponding to the field, the input of the second target model is the output of the first target model, and the output is a field and an attribute value corresponding to the field;

And inputting the work order text into the target model to obtain fields included in the work order text and attribute values corresponding to the fields.

3. The method of claim 1 or 2, further comprising:

detecting a character string selected in the work order text;

determining a field corresponding to the character string, and displaying first prompt information containing the field on the character string;

and detecting clicking the first prompt information, and filling the character string into an input frame corresponding to the field.

4. The method of claim 1 or 2, further comprising:

detecting a search instruction, and determining input search content, wherein the input search content comprises keywords of the work order text;

inputting the input search content into a recommendation model to obtain a first form with similarity meeting requirements with the worksheet text;

determining a search result display rule;

and displaying the obtained first form according to the search result display rule.

5. The method of claim 1 or 2, further comprising:

if the same object query instruction is detected, determining an input object to be queried, wherein the object to be queried comprises a telephone number in the work order text and a theme of the work order text;

Determining a second form with an association relation with the object to be queried, and clustering the second form according to a preset clustering rule;

and displaying the clustering result according to the display rule.

6. The method of claim 1 or 2, further comprising:

if the entering of the signing page is detected, determining a form to be checked, wherein the form to be checked is a form which is completed to be filled based on a work order text;

synchronously auditing the form to be audited from multiple angles by using a call model, and determining character strings which do not accord with preset rules and refusal advice corresponding to the character strings in the form to be audited;

marking the character string in the form to be checked, and displaying second prompt information containing the refusal suggestion corresponding to the character string in a preset display area;

and detecting an instruction for confirming the second prompt information, and filling the refusal sign suggestion corresponding to the second prompt information into the refusal sign reason input box.

7. A business interaction device based on artificial intelligence, comprising:

the system comprises a form filling instruction detection unit, a form filling instruction detection unit and a form filling unit, wherein the form filling instruction detection unit is used for detecting a form filling instruction and determining a work order text and a field to be filled which are input in a form;

the analysis unit is used for calling a model to analyze the work order text to obtain fields contained in the work order text and attribute values corresponding to the fields;

The matching unit is used for matching the field with the field to be filled in, and filling the attribute value corresponding to the field matched with the field to be filled in the input frame corresponding to the field to be filled in;

the display unit is used for displaying the filled form;

the analysis unit analyzes the work order text by calling a model in the following mode to obtain an attribute value corresponding to the incident address:

8. The apparatus of claim 7, wherein the parsing unit is configured to:

9. The apparatus of claim 7 or 8, further comprising: a modifying unit for:

detecting a character string selected in the work order text;

10. The apparatus of claim 7 or 8, further comprising: a search unit for

determining a search result display rule;

11. The apparatus of claim 7 or 8, further comprising: the same object query unit is used for:

and displaying the clustering result according to the display rule.

12. The apparatus of claim 7 or 8, further comprising: the signing and rotating unit is used for:

13. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein, the liquid crystal display device comprises a liquid crystal display device,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

14. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-6.