A proactive perception scheme optimization system based on Agent-based Tt e AI intelligent agent

Abstract

This invention discloses an active perception scheme optimization system based on the Agent-based Tt e AI intelligent agent, relating to the field of voltage control technology. When a user inputs various consultation statements, this invention combines the Agent intelligent agent and historical consultation information to generate consultation statement recommendations corresponding to each input word. When the user does not select a recommended consultation statement, the system performs component analysis and word order correction on each input consultation statement, followed by semantic analysis and emotion classification. Finally, it integrates the semantic and emotion classifications of each consultation statement to match a personalized response to the user, generating a personalized consultation profile for subsequent auxiliary recommendations. This invention, based on Agent-based intelligent perception, makes interaction more efficient and proactive, significantly improving the accuracy of semantic recognition and the relevance of responses, thus achieving intelligent responses to user inquiries.

Description

Technical Field

[0001] This invention relates to the field of AI intelligent perception technology, specifically to an active perception scheme optimization system based on the Agent-mode Tiido AI intelligent agent. Background Technology

[0002] With the deep integration of artificial intelligence technology and customer service, intelligent consultation systems have become a core carrier for various industries to improve service efficiency and optimize user experience. As user consultation needs become more diversified, their expressions more colloquial, and their demands for personalized services increase, traditional intelligent consultation systems are gradually revealing problems such as insufficient response accuracy and a rigid interactive experience. In practical applications, user-input consultation statements often suffer from problems such as non-standard expressions, disordered word order, and semantic ambiguity. Traditional systems have weak word order correction capabilities, only able to correct simple grammatical errors, and cannot accurately analyze components and reconstruct word order for colloquial expressions. This leads to a significant reduction in the accuracy of subsequent semantic analysis and fails to meet the differentiated interactive needs of different users. Therefore, developing an intelligent perception and optimization system is crucial.

[0003] There are some shortcomings in the existing technology in terms of intelligent perception, which are reflected in the following aspects: (1) Most of the existing intelligent consultation systems passively receive the user's complete consultation statement and then process it. They cannot capture the input words and generate consultation statement recommendations in real time during the user's input process. They cannot shorten the user's input path, nor can they predict the user's core needs in advance, resulting in cumbersome user interaction operations and insufficient initiative in service response.

[0004] (2) When constructing a consultation statement database, existing technologies often directly use the raw data of historical consultation statements without decomposing the grammatical components and optimizing semantic deduplication. This results in a large amount of redundant data with semantic repetition in the statement database. When performing similarity matching, redundant data not only reduces matching efficiency but also affects matching accuracy, making it impossible to provide users with accurate statement recommendations.

[0005] (3) When constructing a consultation statement database, existing technologies mostly use the raw data of historical consultation statements directly without decomposing the grammatical components and optimizing semantic deduplication. This results in a large amount of redundant data with semantic repetition in the statement database. When performing similarity matching, redundant data not only reduces matching efficiency but also affects matching accuracy, making it impossible to provide users with accurate statement recommendations. At the same time, the semantic matching of existing systems mostly relies on surface keyword comparison and does not introduce a deep semantic matching model, making it impossible to capture the deep semantic relationships of statements. In addition, the matching process does not combine grammatical adaptability and scenario fit for comprehensive judgment, making it difficult to select the most consistent answer content with the user's consultation logic, which easily leads to irrelevant answers. Summary of the Invention

[0006] To address the aforementioned technical shortcomings, the purpose of this invention is to provide an optimized system for an active perception scheme based on the Agent-based Tt e AI intelligent agent.

[0007] To solve the above technical problems, the present invention adopts the following technical solution: The present invention provides an active perception scheme optimization system based on the Agent-mode Tiido AI intelligent agent, including: an input correction module: used to obtain each input word of the user in real time when the user inputs each consultation statement, and generate consultation statement recommendations corresponding to each input word by combining the Agent intelligent agent and historical consultation information.

[0008] The word order correction module is used to analyze the components of each consultation statement entered by the user when the user has not selected any of the recommended consultation statements, and then correct the word order of each consultation statement entered by the user.

[0009] Semantic and Emotion Perception Module: This module performs semantic analysis on each consultation statement and classifies them according to their emotion. It then combines the semantic and emotion classifications of each consultation statement to obtain a personalized response for the user.

[0010] User profile analysis module: Used to determine whether to generate a personalized consultation profile for a user, and to provide an auxiliary recommendation process based on the user's personalized consultation profile after the response.

[0011] The beneficial effects of the present invention are as follows: (1) The present invention relies on the real-time perception capability of the Agent intelligent body to actively acquire words and generate sentence recommendations during the user's input of consultation sentences. Then, it combines historical consultation information to construct a deduplicated and optimized reference sentence library, thereby realizing the advance prediction of user needs. This not only shortens the user input path and reduces the operational costs caused by colloquial and non-standard input, but also reduces the difficulty of user input through accurate recommendations. It breaks the limitation of the traditional system passively waiting for complete input, making the interaction more efficient and proactive, significantly improving the user interaction experience, and realizing the intelligent response to user consultations.

[0012] (2) This invention constructs a collaborative closed loop of four major modules: input correction, word order correction, semantic and emotion perception, and user profiling. Data from each module is shared in real time and results are deeply linked. The standardized sentences output by the word order correction module provide accurate input for semantic analysis. The semantic parsing and emotion perception results are integrated to generate personalized responses, which effectively solves the problems of traditional module fragmentation and superficial semantic matching, greatly improves the accuracy of semantic recognition and the relevance of responses, and avoids irrelevant answers.

[0013] (3) This invention constructs an emotion set by using emotional keywords and tone-enhancing words, and establishes a precise mapping system between negative emotions and exclusive comforting words. It provides differentiated empathetic comfort for different negative emotions, and then connects with core consultation responses, realizing a dual service of emotional comfort and problem-solving. Compared with the problem of disconnect between emotional perception and response in traditional systems, this invention makes the service warmer, effectively alleviates users' negative emotions, and improves users' satisfaction and recognition of the service.

[0014] (4) This invention constructs a multi-dimensional personalized profile based on users' historical consultation data, and provides accurate auxiliary recommendations or service guarantees after the response, combining the profile with the current consultation scenario. It not only breaks the traditional homogeneous service model and meets the differentiated needs of different users, but also continuously optimizes service strategies through dynamic profile updates, providing users with full-cycle service empowerment. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the system structure connection of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Reference Figure 1 As shown, the present invention provides an active perception scheme optimization system based on the Agent-mode Tiido AI intelligent agent, including an input correction module: used to obtain each input word of the user in real time when the user inputs each consultation statement, and combine the Agent intelligent agent and historical consultation information to generate consultation statement recommendations corresponding to each input word.

[0019] It should be noted that by actively capturing the input words and quickly generating accurate consultation recommendations during the real-time process of user input, user needs can be anticipated in advance, the user input path can be shortened, and the proactive service awareness of the intelligent agent can be strengthened.

[0020] In a specific example, when a user inputs each consultation statement, the system acquires each input word in real time, combines it with the Agent and historical consultation information, and generates consultation statement recommendations corresponding to each input word. The specific process is as follows: First, the Agent dynamically records each input word of the user in real time. At the same time, it calls the Web database interface to query each synonym, synonym expression, and common variant of each input word of the user. The user's input words and their corresponding synonyms, synonym expressions, and common variants are combined and recorded as the user's target consultation set.

[0021] It should be noted that the query provides synonyms, related expressions, and common variations for each word entered by the user. For example, for "how", the synonym for "how" is "how"; for "balance", the synonym for "remaining quota" is "remaining quota". For related expressions, for example, for "Why hasn't the order been shipped yet", the synonym for "Why hasn't the order been shipped yet, what is the reason?" is "Why hasn't the order been shipped yet, what is the reason?" is "Why hasn't the order been shipped yet, what is the reason?" is "Why hasn't the broadband been cancelled?" is "How to cancel broadband", and common variations, for example, common variations of "How to cancel broadband" are "broadband cancellation process", etc.

[0022] The agent retrieves historical consultation information from the data center and constructs a consultation reference statement library based on this information. Then, using a semantic similarity algorithm, it compares the target consultation set with each historical consultation statement recorded in the consultation reference statement library, outputs the similarity between the target consultation set and each historical consultation statement in the library, sorts the similarity scores from highest to lowest, and selects several corresponding historical consultation statements in descending order of similarity scores, which are then pushed to the user's consultation end.

[0023] In a specific example, the agent retrieves historical consultation information from the data center of the consultation platform and constructs a consultation reference statement library based on this information. The specific construction process is as follows: Based on the agent, historical consultation statements within a preset historical time period are retrieved from the data center. First, each historical consultation statement is preprocessed. Then, a syntax parsing engine is used to decompose the grammatical components of each historical consultation statement, extracting the subject, predicate, object, attributive, adverbial, and complement. The subject of any historical consultation statement is designated as the target subject. Using a semantic similarity algorithm, the subjects and target subjects corresponding to the remaining historical consultation statements are output. The similarity between the subject and the target subject of each historical consultation statement is compared with a set semantic similarity threshold. Subjects of other historical consultation statements with similarity greater than or equal to the set semantic similarity threshold are then deleted. This process is repeated until only one subject exists for each semantic category. The predicate, object, attributive, adverbial, and complement of each historical consultation statement are then processed in the same way. Finally, the processed subjects, predicates, objects, attributives, adverbials, and complements of each historical consultation statement are arbitrarily combined according to semantic logic to construct standard consultation statements. These standard consultation statements are then used to construct the consultation reference statement library for the consultation platform.

[0024] It should be noted that each historical consultation statement is preprocessed, including deduplication and removal of invalid statements.

[0025] It should also be noted that the semantic similarity threshold is set by the relevant staff and no specific restrictions are imposed here.

[0026] The word order correction module is used to analyze the components of each consultation statement entered by the user when the user has not selected any of the recommended consultation statements, and then correct the word order of each consultation statement entered by the user.

[0027] It should be noted that performing grammatical analysis and word order optimization on non-standard user input can eliminate ambiguity and ensure the accuracy of subsequent semantic and emotion perception.

[0028] In a specific example, the process of parsing the components of each consultation statement input by the user and then correcting the word order of each consultation statement input by the user is as follows: First, the agent removes the modal particles from each consultation statement input by the user. Then, the grammatical components of each consultation statement after removing the modal particles are identified and labeled. Based on this, the subject, predicate, object, attributive, adverbial and complement of each consultation statement are identified. Then, the subject, predicate and object of each consultation statement are arranged and combined in the order of subject + predicate + object to obtain the standard consultation statement input by the user.

[0029] Semantic and Emotion Perception Module: This module performs semantic analysis on each consultation statement and classifies them according to their emotion. It then combines the semantic and emotion classifications of each consultation statement to obtain a personalized response for the user.

[0030] In a specific example, the semantic analysis of each consultation statement is performed as follows: First, based on the Agent, the syntactic components of each standard consultation statement are matched with the set consultation intent keyword library to identify the user's consultation intent. Then, based on the syntactic components of each standard consultation statement input by the user, the subject object and query requirements corresponding to the consultation intent are identified.

[0031] The Agent matches hierarchical scenario tags to the corresponding query statements based on the query intent, the subject, and the query requirements. The hierarchical scenario tags adopt a three-level structure: the first-level scenario tag corresponds to the user's core query intent, the second-level scenario tag corresponds to the subject, and the third-level scenario tag corresponds to the query requirements.

[0032] In a specific example, the process of classifying the emotions of each consultation statement is as follows: using an emotion keyword recognition model, each emotion keyword and each tone-enhancing word is extracted from each original consultation statement input by the user. The extracted emotion keywords and tone-enhancing words are then combined to construct the user's emotion keyword set. The user's emotion keyword set is then matched with a set of set emotion classification labels to obtain the emotion classification label of the user's current consultation. The emotion classification labels include positive emotion labels, negative emotion labels, and neutral emotion labels.

[0033] It should be noted that emotional keywords include "satisfaction," "disappointment," "anger," and "rage," while intensifying words include "very," "simply," and "extremely."

[0034] In a specific example, the process of combining the semantic and emotional classifications of each consultation statement to obtain a personalized response for the user is as follows: When the Agent identifies that the emotional classification label of the user's current consultation is a positive or neutral emotional label, it obtains the corresponding consultation response based on the hierarchical scenario label analysis of each consultation statement.

[0035] When the agent identifies that the current user's emotional category tag is negative, it matches the corresponding soothing response based on the user's corresponding emotional keyword set, and then obtains the corresponding consultation response based on the hierarchical scenario tags of each user's consultation statement. The agent then uses both soothing and consultation responses to reply to the user.

[0036] It should be noted that the Agent intelligent agent builds its dialogue system based on negative emotion types. Each type of negative emotion corresponds to exclusive core keywords, tone reinforcement words, and exclusive soothing responses. For example, the core keywords for anger are "angry, furious, and complaining," the corresponding tone reinforcement words are "extremely," "simply," and "excessive," and the corresponding exclusive soothing responses are: "We are very sorry for the bad experience you had. We will immediately check the problem and handle it properly for you." Based on this, the agent matches the corresponding soothing responses according to the user's set of emotional keywords.

[0037] In a specific example, the corresponding consultation response is obtained by analyzing the hierarchical scenario tags of each user's consultation statement. The specific analysis process is as follows: First, a hierarchical scenario tag framework for consultation responses is constructed, which includes first-level tags, second-level tags, and third-level tags. The first-level scenario tag corresponds to the consultation intent response, the second-level scenario tag corresponds to the subject object response, and the third-level scenario tag corresponds to the query requirement response. The responses corresponding to the first-level tags, second-level tags, and third-level tags are recorded as standard consultation responses.

[0038] Then, retrieve several customer service replies from the data center, break down each reply into its components, identify the modifiers, adverbs, and complements in each reply, and record the corresponding modifiers, adverbs, and complements in each reply as supplementary consultation replies.

[0039] The standard consultation response is arbitrarily matched with each supplementary consultation response to obtain each alternative consultation response. A relevant matching model is then used to match each alternative consultation response with the consultation statement entered by the user to obtain the matching degree between the alternative consultation response and the consultation statement entered by the user. The alternative consultation response with the highest matching degree is taken as the corresponding consultation response and output to the user.

[0040] It should be noted that a relevance matching model is used to match each alternative consultation response with the user's input consultation statement. The relevance matching model includes U... 2 The U-IMN model, Sentence-BERT model, and consistency parsing model are all existing technologies and will not be elaborated further.

[0041] User profile analysis module: Used to determine whether to generate a personalized consultation profile for a user, and to provide an auxiliary recommendation process based on the user's personalized consultation profile after the response.

[0042] In a specific instance, the process of determining whether to generate a personalized consultation profile for a user is as follows: obtain the user's total number of historical consultations from the data center, compare the user's total number of consultations with a set consultation number threshold, and generate the user's personalized consultation profile when the user's total number of historical consultations reaches the set consultation number threshold.

[0043] It should be noted that the threshold for the number of inquiries is set by the relevant staff, and will not be elaborated here.

[0044] In a specific example, the auxiliary recommendation process after responding to a user's personalized consultation profile is as follows: First, the agent retrieves the user's historical consultation information from the data center. Based on this information, it obtains the number of times the user searched for consultation intent and subject after each consultation, and then calculates the total number of relevant consultation searches. The total number of relevant consultation searches is compared with a set search threshold. If the total number of relevant consultation searches is greater than or equal to the threshold, the user is recorded as a recommended user for the target business; otherwise, they are recorded as an unrecommended user for the target business.

[0045] It should be noted that the search frequency threshold is set by the relevant staff and is not specifically limited here.

[0046] After completing the response to the current inquiry, the Agent intelligent agent calls upon the user's personalized inquiry profile. Combining the inquiry intent and the target audience, if the user is a recommended user for the target business, relevant business information will be recommended after the inquiry response; if the user is not a recommended user for the target business, no relevant business recommendations will be made.

[0047] The examples described in this invention are not limited to the specific embodiments listed above. The examples are merely illustrative to facilitate understanding of the invention and do not constitute a limitation on the scope of protection of this invention. Any modifications, equivalent substitutions, etc., made within the spirit and principles of this invention should be included within the scope of protection.

[0048] The above description is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they should all fall within the protection scope of the present invention.

Claims

1. An active perception scheme optimization system based on the Agent-based Tt e AI intelligent agent, characterized in that, Includes the following modules: Input correction module: It is used to obtain the user's input words in real time when the user inputs each consultation statement, and combine them with the Agent and historical consultation information to generate consultation statement recommendations corresponding to each input word; Sentence order correction module: When the user does not select any of the recommended consultation statements, it performs component analysis on each consultation statement entered by the user, and then corrects the word order of each consultation statement entered by the user. Semantic and Emotion Perception Module: This module performs semantic analysis on each consultation statement and classifies each consultation statement according to its emotion. It then combines the semantic and emotion classifications of each consultation statement to match and obtain a personalized response for the user. User profile analysis module: Used to determine whether to generate a personalized consultation profile for a user, and to provide an auxiliary recommendation process based on the user's personalized consultation profile after the response.

2. The active perception scheme optimization system based on the Agent-mode Tt e AI intelligent agent according to claim 1, characterized in that, As the user inputs each query statement, the system acquires each input word in real time, combines it with the agent and historical query information to generate query statement recommendations corresponding to each input word. The specific process is as follows: First, the Agent intelligent agent dynamically records each word input by the user in real time. At the same time, it calls the Web database interface to query each word input by the user, and queries each synonym, synonym expression and common variant. The user's words input by the user and their corresponding synonyms, synonym expressions and common variants are combined and recorded as the user's target consultation set. The agent retrieves historical consultation information from the data center and constructs a consultation reference statement library based on this information. Then, using a semantic similarity algorithm, it compares the target consultation set with each historical consultation statement recorded in the consultation reference statement library, outputs the similarity between the target consultation set and each historical consultation statement in the library, sorts the similarity scores from highest to lowest, and selects several corresponding historical consultation statements in descending order of similarity scores, which are then pushed to the user's consultation end.

3. The active perception scheme optimization system based on the Agent-mode Tt e AI intelligent agent according to claim 2, characterized in that, The agent retrieves historical consultation information from the data center and constructs a consultation reference statement library for the consultation platform based on this information. The specific construction process is as follows: Based on an agent-based intelligent system, historical query statements within a preset historical time period are retrieved from the data center. First, each historical query statement is preprocessed. Then, a syntax parsing engine decomposes the grammatical components of each historical query statement, extracting the subject, predicate, object, attributive, adverbial, and complement. The subject of any historical query statement is designated as the target subject. A semantic similarity algorithm is used to output the similarity between the subject of each other historical query statement and the target subject. Finally, the similarity between the subject of each other historical query statement and the target subject is compared with the set semantic similarity. The similarity threshold is compared, and then the subjects corresponding to the remaining historical consultation statements with similarity greater than or equal to the set semantic similarity threshold are deleted. The above steps are repeated until there is only one corresponding subject for each semantic category. The predicate, object, attributive, adverbial and complement of each historical consultation statement are processed in the above manner. Finally, the subjects, predicates, objects, attributives, adverbials and complements corresponding to each historical consultation statement are arbitrarily combined according to semantic logic to construct each standard consultation statement. Then, the consultation reference statement library of the consultation platform is constructed from each standard consultation statement.

4. The active perception scheme optimization system based on the Agent-mode Tt e AI intelligent agent according to claim 3, characterized in that, The process of parsing the components of each query statement input by the user and then correcting the word order of each query statement input by the user is as follows: First, the agent removes the modal particles from each query statement input by the user. Then, the grammatical components of each query statement after removing the modal particles are identified and labeled. Based on this, the subject, predicate, object, attributive, adverbial and complement of each query statement are identified. Then, the subject, predicate and object of each query statement are arranged and combined in the order of subject + predicate + object to obtain the standard query statement input by the user.

5. The active perception scheme optimization system based on the Agent-mode Tt e AI intelligent agent according to claim 4, characterized in that, The semantic analysis of each consultation statement is performed as follows: First, the Agent intelligent body matches the grammatical components of each standard consultation statement with the set consultation intent keyword library to identify the user's consultation intent. Then, based on the grammatical components of each standard consultation statement input by the user, the subject object and query requirements corresponding to the consultation intent are identified. The Agent matches hierarchical scenario tags to the corresponding query statements based on the query intent, the subject, and the query requirements. The hierarchical scenario tags adopt a three-level structure: the first-level scenario tag corresponds to the user's core query intent, the second-level scenario tag corresponds to the subject, and the third-level scenario tag corresponds to the query requirements.

6. The active perception scheme optimization system based on the Agent-mode Tt e AI intelligent agent according to claim 5, characterized in that, The process of classifying the emotions of each consultation statement is as follows: The emotional keyword recognition model extracts emotional keywords and tone-enhancing words from the original consultation statements input by the user. The extracted emotional keywords and tone-enhancing words are combined to construct the user's emotional keyword set. Then, the user's emotional keyword set is matched with the set of emotional classification labels to obtain the emotional classification label of the user's current consultation. The emotional classification labels include positive emotional labels, negative emotional labels, and neutral emotional labels.

7. The active perception scheme optimization system based on the Agent-mode Tt e AI intelligent agent according to claim 6, characterized in that, The process of combining the semantic and emotional classifications of each inquiry statement to obtain a personalized response for the user is as follows: When the agent identifies that the current emotional category label of the user's inquiry is a positive emotional label or a neutral emotional label, it obtains the corresponding inquiry response based on the hierarchical scenario label analysis of each inquiry statement. When the agent identifies that the current user's emotional category tag is negative, it matches the corresponding soothing response based on the user's corresponding emotional keyword set, and then obtains the corresponding consultation response based on the hierarchical scenario tags of each user's consultation statement. The agent then uses both soothing and consultation responses to reply to the user.

8. The active perception scheme optimization system based on the Agent-mode Tt e AI intelligent agent according to claim 7, characterized in that, The corresponding consultation response is obtained by analyzing the hierarchical scenario tags of each user's consultation statement. The specific analysis process is as follows: First, a hierarchical scenario-based tagging framework for consultation responses is constructed. The hierarchical scenario tags for consultation responses include first-level tags, second-level tags, and third-level tags. The first-level scenario tags correspond to the consultation intent response, the second-level scenario tags correspond to the subject object response, and the third-level scenario tags correspond to the query requirement response. The responses corresponding to the first-level tags, second-level tags, and third-level tags are recorded as standard consultation responses. Then, retrieve several customer service replies from the data center, break down each reply into its components, identify the modifiers, adverbs, and complements in each reply, and record the corresponding modifiers, adverbs, and complements in each reply as supplementary consultation replies. The standard consultation response is arbitrarily matched with each supplementary consultation response to obtain each alternative consultation response. A relevant matching model is then used to match each alternative consultation response with the consultation statement entered by the user to obtain the matching degree between the alternative consultation response and the consultation statement entered by the user. The alternative consultation response with the highest matching degree is taken as the corresponding consultation response and output to the user.

9. The active perception scheme optimization system based on the Agent-mode Tt e AI intelligent agent according to claim 8, characterized in that, The specific process for determining whether to generate a personalized consultation profile for the user is as follows: The system retrieves the user's total historical consultation count from the data center and compares it with a set consultation count threshold. When the user's total historical consultation count reaches the set consultation count threshold, a personalized consultation profile for the user is generated.

10. The active perception scheme optimization system based on the Agent-mode Tt e AI intelligent agent according to claim 9, characterized in that, The auxiliary recommendation process based on the user's personalized consultation profile and subsequent response is as follows: First, the agent retrieves the user's historical consultation information from the data center. Based on this information, it obtains the number of times the user searched for the consultation intent and the subject after each consultation. Then, it calculates the total number of relevant consultation searches for the user and compares it with a set search threshold. If the total number of relevant consultation searches is greater than or equal to the threshold, the user is recorded as a recommended user for the target business; otherwise, the user is recorded as an unrecommended user for the target business. After completing the response to the current inquiry, the Agent intelligent agent calls up the user's personalized inquiry profile, and combines the inquiry intent and the main target of the inquiry. If the user is a target business user, relevant business information recommendations are attached after the inquiry response. If a user is not eligible for recommendation by the target business, then no related business recommendations will be made.

Images