WO2024057301A1

WO2024057301A1 - System and computer implemented method for personalized nurturing of prospective real-estate clients

Info

Publication number: WO2024057301A1
Application number: PCT/IL2023/050965
Authority: WO
Inventors: Erez Yakoel; Asaf Rubin
Original assignee: Rosetal System Information Ltd
Priority date: 2022-09-15
Filing date: 2023-09-07
Publication date: 2024-03-21

Abstract

Disclosed are a computer implemented method and system for personalized nurturing of a prospective real-estate client, using language processing, big data analysis and machine learning algorithms.

Description

SYSTEM AND COMPUTER IMPLEMENTED METHOD FOR PERSONALIZED NURTURING OF PROSPECTIVE REAL-ESTATE CLIENTS

FIELD OF THE INVENTION

Embodiments of the disclosure relate to automated nurturing of prospective clients, in particular to machine learning algorithm-based computational methods for clustering of clients and/or for computing a personalized customer relationship management (CRM) scheme.

BACKGROUND

Currently, real-estate agents receive a multiplicity of leads. However, due to the large number of leads, the agents are typically unable to approach each of them and many of the leads remain unanswered.

Moreover, real-estate agents are typically unable to up-front predict on which clients, when and how a follow-up interaction should be conducted in order to increase the likelihood of transaction.

There is, therefore, a need for a system and method capable of reliably clustering prospective clients based on their real-estate associated behavior and/or for deriving a personalized customer relationship management (CRM) scheme.

SUMMARY OF THE EMBODIMENTS

According to some embodiments, there is provided a computer implemented method and/or platform for personalized nurturing of prospective real-estate clients.

Advantageously, the herein disclosed method and system applies Al models, such as, but not limited to, natural language processing (NLP) and machine learning (ML), on the digital behavior data and features extracted therefrom of a multiplicity of real-estate leads, in order to increase their likelihood of transaction. As a further advantage, by predicting a most efficient next best action for the multiplicity of leads, the likelihood of transaction for each of the leads may be increased.

Moreover, by automating at least a portion of the interaction with the leads, real estate agents are able to devote their time and resources to mature leads only. In addition, the hereindisclosed method and platform enables processing a large plurality of leads, thereby ensuring that a prospective client that is mature for conducting a real estate interaction is not disregarded or neglected.

Advantageously, the hereindisclosed computer implemented method and platform provides personalized automation and assistance with real-estate transactions, which are often the biggest transaction in the life of subjects and involves complex decision making that includes objective as well as subjective factors.

According to some embodiments, there is provided a computer implemented method for personalized nurturing of prospective real-estate clients, the method comprising: a) obtaining a prospective real-estate client (and optionally from a large plurality of clients); b) obtaining and/or extracting data concerning the client; c) applying one or more machine learning modules on the extracted data to determine an initial real-estate preference associated with the client; d) conducting one or more digital interactions with the client, and extracting from the interaction: data regarding at least one digital interaction feature of the client; wherein the at least one digital interaction feature is selected from, an amount of time spent browsing per property, a quantity of property listings checked, a degree of similarity between the properties checked, a degree of consistency between the property listings checked, an amount of sharing of the checked property listing or any combination thereof; and a user feedback (text/voice) selected from a response of the prospective client to a content of the digital interaction, a sentiment of the prospective client with respect to the content of the digital interaction, a request of the prospective client with regards to future digital interaction content and any combination thereof, wherein extracting the user feedback comprises applying one or more NLP models on the digital interaction, e) processing the extracted initial real-estate preference, the at least one digital interaction feature and the user feedback, by applying one or more machine learning modules thereon; wherein the processing comprises computing a likelihood of transaction for the client based on the processed initial real-estate preference, on the at least one digital interaction feature extracted from the one or more interactions; f) computing a next best action (NBA) for the prospective client, based on the processing, wherein the NBA is configured to increase the likelihood of transaction of the prospective client, wherein the NBA is selected from: asking for more information or clarification on preferences of the prospective client, suggesting an open house, notifying the prospective client on changes in market, asking an advisor to contact the client and any combination thereof, wherein a time, a type and a content of the NBA is determined based on the processed initial real-estate preference, digital interaction feature and user feedback; and g) automatically executing the NBA, thereby increasing the likelihood of transaction of the prospective client, wherein the NBA is a machine -based interaction.

According to some embodiments, the method further includes repeating steps d-g until the likelihood to transact exceeds a predetermined threshold value. According to some embodiments, the method further includes repeating steps d-g until the likelihood to transact exceeds 0.5%, 1%, 5% or 10%. Each possibility is a separate embodiment.

According to some embodiments, the method further includes extracting, from the executed NBA, additional digital interaction features, and reapplying a second machine learning module on the additional digital interaction features, thereby refining the computed likelihood of transaction.

According to some embodiments, the extracting data concerning the client includes automatically extracting from pages that can be accessed by a Web browser. According to some embodiments, the method further includes calculating a point in time when a specific likelihood to engage in the transaction is reached.

According to some embodiments, the method further includes matching an agency and/or an agent to the client, based on the client type cluster into which the client is classified.

According to some embodiments, the type of the NBA is selected from: texting, voice messaging, image messaging, multimedia messaging, bot conversations or any combination thereof. Each possibility is a separate embodiment.

According to some embodiments, the content of the interaction comprises listings, image quality, language, language level, or any combination thereof. Each possibility is a separate embodiment.

According to some embodiments, the method further includes determining one or more unique preferences of the client. According to some embodiments, the time, the type and the content of the NBA is further determined based on the one or more unique preferences.

According to some embodiments, the method further includes relaxing the real-estate preference and/or the one or more unique preferences, when resulting in a low inventory.

According to some embodiments, the method further includes identifying one or more blockers to transaction, based on the client type cluster and on the one or more unique preferences. According to some embodiments, the NBA comprises a resolution to the identified one or more blockers.

According to some embodiments, the real-estate preference comprises one or more of: buying/selling of a property, renting/leasing a property and/or financing a property, acquiring a real-estate related legal service, a real-estate related insurance related service, renovation service, moving service, a real-estate related financial service or any combination thereof. Each possibility is a separate embodiment.

According to some embodiments, the first and second machine learning modules comprise one or more of a voice analysis model, text analysis model, a voice to text transcription model, a transcript analysis model, a sentiment analysis model, an image analysis model or any combination thereof. Each possibility is a separate embodiment.

According to some embodiments, the client is a subject visiting, entering and/or communicating with a virtual real estate platform.

According to some embodiments, there is provided a system for personalized nurturing of a prospective real-estate client, the system comprising a processor configured to: obtain and/or extract data concerning the prospective client; apply one or more machine learning modules on the extracted data to determine an initial real-estate preference associated with the prospective client; extract from one or more interactions conducted with the prospective client: data regarding at least one digital interaction feature of the prospective client; wherein the at least one digital interaction feature is selected from an amount of time spent browsing per property, a quantity of property listings checked, a degree of similarity between the properties checked, a degree of consistency between the property listings checked, an amount of sharing of the checked property listing or any combination thereof; and a user feedback (text/voice) selected from a response of the prospective client to a content of the digital interaction, a sentiment of the prospective client with respect to the content of the digital interaction, a request of the prospective client with regards to future digital interaction content and any combination thereof, wherein extracting the user feedback comprises applying one or more NLP models on the digital interaction; process the extracted initial real-estate preference, the at least one digital interaction feature and the user feedback, by applying one or more machine learning modules thereon; wherein the processing comprises computing a likelihood of transaction for the client based on the processed initial real-estate preference, on the at least one digital interaction feature extracted from the one or more interactions; compute a next best action (NBA) for the prospective client, based on the processing, wherein the NBA is configured to increase the likelihood of transaction of the prospective client, wherein the NBA is selected from asking for more information or clarification on preferences of the prospective client, suggesting an open house, notifying the prospective client on changes in market, asking an advisor to contact the client and any combination thereof, wherein a time, a type and a content of the NBA is determined based on the processed initial real-estate preference, digital interaction feature and user feedback; and automatically execute the NBA, thereby increasing the likelihood of transaction of the prospective client, wherein the NBA is a machine -based interaction.

According to some embodiments, there is provided a computer implemented method for personalized nurturing of prospective real-estate clients, the method comprising: identifying a prospective client; obtaining and/or extracting data concerning the client; applying one or more machine learning modules on the extracted data to determine an initial real-estate preference associated with the prospective client; conducting one or more interactions with the prospective client, and extracting from the interaction data regarding at least one digital interaction feature of the prospective client; applying the one or more machine learning modules on the initial real-estate preference, and on the at least one digital interaction feature extracted from the one or more interactions, to classify the prospective client into a client type cluster; and computing a customer relationship management (CRM) scheme for the prospective client, based on the client type cluster, wherein the CRM scheme comprises initiating at least one subsequent interaction with the prospective client, wherein a time, a type and a content of the subsequent interaction is determined, based on the client type cluster into which the prospective client is classified.

According to some embodiments, the method further comprises extracting, from the at least one subsequent interaction, additional digital interaction features, and reapplying the second machine learning module on the additional digital interaction features, thereby refining the classification of the prospective client.

According to some embodiments, the method further comprises applying a third machine learning module to compute a likelihood of transaction for the prospective client. According to some embodiments, the second machine learning module is further applied on the computed likelihood of transaction.

According to some embodiments, the extracting data concerning the client comprises automatically extracting from pages that can be accessed by a Web browser.

According to some embodiments, the method further comprises calculating a point in time when a specific likelihood to engage in the transaction is reached.

According to some embodiments, the method further comprises matching an agency and/or an agent to the prospective client, based on the client type cluster into which the prospective client is classified.

According to some embodiments, the type of the first and the additional interaction is selected from: texting, voice messaging, image messaging, multimedia messaging, bot conversations, voice calls, video calls or any combination thereof. Each possibility is a separate embodiment.

According to some embodiments, the content of the interaction comprises type of listings, type of images, language, language level, or any combination thereof, based on the determined communication and real-estate preferences. Each possibility is a separate embodiment.

According to some embodiments, the method further comprises determining one or more unique preferences of the prospective client. According to some embodiments, the time, the type and the content of the subsequent interaction is further determined based on the one or more unique preferences.

According to some embodiments, the method further comprises relaxing the real-estate preference and/or the one or more unique preferences, when resulting in a low inventory.

According to some embodiments, the method further comprises identifying one or more blockers to transaction, based one the client type cluster and on the one or more unique preferences. According to some embodiments, the content of the at least one subsequent interaction comprises a resolution to the identified one or more blockers.

According to some embodiments, the initiated additional interaction is a machine -based interaction.

According to some embodiments, the prospective client is a subject visiting, entering and/or communicating with a virtual real estate platform. Each possibility is a separate embodiment.

According to some embodiments, the at least one digital interaction feature is selected from an amount of time spent browsing, a degree to which the browsing was an in-depth browsing, a quantity of property listings checked, a degree of similarity between the properties checked, a degree of consistency between the property listings checked and a set of preferences investigated by the prospective client, a listing checking rate, a realistic expectations index, a level of actiontaking related to a potential real estate transaction, an amount of people involved by the prospective client with regards to the property listing, an amount of sharing of the checked property listing or any combination thereof. Each possibility is a separate embodiment.

According to some embodiments, there is provided a system for personalized nurturing of a prospective real-estate client, the system comprising a processor configured to: obtain and/or extract data concerning the prospective client; apply one or more machine learning modules on the extracted data to determine an initial real-estate preference associated with the prospective client; extract from one or more interactions conducted with the prospective client data regarding at least one digital interaction feature of the prospective client; apply the one or more machine learning modules on the initial real-estate preference, and on the at least one digital interaction feature extracted from the one or more interactions, to classify the prospective client into a client type cluster; computing a next best action (NBA) for the prospective client, thereby increasing the likelihood that the subject will conduct a real-estate interaction.

Certain embodiments of the present disclosure may include some, all, or none of the above advantages. One or more other technical advantages may be readily apparent to those skilled in the art from the figures, descriptions, and claims included herein. Moreover, while specific advantages have been enumerated above, various embodiments may include all, some, or none of the enumerated advantages.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. In case of conflict, the patent specification, including definitions, governs. As used herein, the indefinite articles “a” and “an” mean “at least one” or “one or more” unless the context clearly dictates otherwise.

BRIEF DESCRIPTION OF THE FIGURES

Some embodiments of the disclosure are described herein with reference to the accompanying figures. The description, together with the figures, makes apparent to a person having ordinary skill in the art how some embodiments may be practiced. The figures are for the purpose of illustrative description and no attempt is made to show structural details of an embodiment in more detail than is necessary for a fundamental understanding of the disclosure. For the sake of clarity, some objects depicted in the figures are not drawn to scale. Moreover, two different objects in the same figure may be drawn to different scales. In particular, the scale of some objects may be greatly exaggerated as compared to other objects in the same figure. In block diagrams and flowcharts, certain steps may be conducted in the indicated order only, while others may be conducted before a previous step, after a subsequent step or simultaneously with another step. Such changes to the orders of the step will be evident for the skilled artisan. Chat bot conversations are indicated in balloons and user instructions provided through selecting an icon or an option from a scroll down menu is indicated by grey boxes. It is understood that combining both text conversations and buttons is optional, and that the entire conversation tree may be through text messages or even, but generally less preferred, through instruction buttons and/or scroll-down menus.

FIG. 1 schematically shows an outline of the herein disclosed Al-based platform for personalized nurturing of prospective real-estate clients, according to some embodiments;

FIG. 2A is an exemplary flowchart of a computer implemented method for generating a personalized nurturing of prospective real-estate clients, according to some embodiments;

FIG. 2B is an exemplary flowchart of a computer implemented method for generating a personalized nurturing of prospective real-estate clients including a feed-back loop, according to some embodiments;

FIG. 3 schematically shows an outline of the herein disclosed Al-based next best action platform for personalized nurturing of prospective real-estate clients, according to some embodiments;

FIG. 4 is an exemplary nurturing conducted on 6,500 dead leads using the herein disclosed system and method.

DETAILED DESCRIPTION OF THE EMBODIMENTS

The following detailed description is of the best currently contemplated modes of carrying out the invention. The description is not to be taken in a limiting sense, but is made merely for the purpose of illustrating the general principles of the invention, since the scope of the invention is best defined by the appended claims.

According to some embodiments, there is provided a computer implemented method and/or platform for personalized nurturing of prospective real-estate clients. As used herein, the terms, “prospective client” and “lead” may be used interchangeably and may refer to any subject that has shown interest in conducting a real-estate transaction. According to some embodiments, the prospective client may be a subject that has visited, entered and/or communicated with a virtual real estate platform, such as, but not limited to, a real -estate website, a real-estate social media page, a real-estate chat-bot or the like. According to some embodiments, the prospective client may be a subject that has contacted a real-estate agent, optionally regarding a specific property listing.

As used herein, the term “nurturing” may refer to steps made in order to promote, stimulate, and/or develop engagement of a prospective client in a real-estate transaction. According to some embodiments, the nurturing may include approaching the prospective client, preferably in an automated manner (e.g. by sending real-estate content to the prospective client). Additionally or alternatively, the nurturing may include interacting with the prospective client, preferably in an automated manner (e.g. via chat bot) in order to increase the engagement of the prospective client and/or the likelihood of transaction.

As used herein, the term “personalized” with regards to the nurturing, may refer to computing a next best action (NBA) that will increase the likelihood that the prospective client will conduct a real-estate transaction. Non-limiting examples of suitable NBAs include: asking for more information or clarification on his/her preferences, suggesting an open house, notifying the client on changes in market (e.g. price changed, lower/higher inventory, market trends, etc), asking an advisor to contact the client and any combination thereof. Each possibility is a separate embodiment. According to some embodiments, computing the NBA comprises automatically executing the NBA. According to some embodiments, automatically executing the NBA comprises exiting the NBA at a time determined as optimal for the prospective client. According to some embodiments, automatically executing the NBA comprises exiting the NBA via a media (e.g. email, sms, video call etc.) determined as optimal for the prospective client.

As used herein, the terms “property listing” and “real estate listing” may be used interchangeably and may refer to any printed advertisement, internet posting, or publicly displayed sign of properties/real estate, which are available for purchasing and/or rent.

As used herein, the terms “real-estate property” and “property” may be used interchangeably and may refer to a single property or a project. Non-limiting examples of single properties include a house/home, an apartment, an office, a fabric, a storage, a land, a commercial building, an industrial property, an agricultural property, a mixed-use property and the like. Nonlimiting examples of real-estate projects include a housing complex, a commercial area, a hotel, and the like.

As used herein the terms “real-estate transaction” and “transaction” may be used interchangeably and may refer to the buying/selling of a property or a part thereof, renting/leasing a property or a part thereof, house-swapping, financing a property or a part thereof and/or investing in a property or a part thereof, real-estate related service or any combination thereof. Each possibility is a separate embodiment. According to some embodiments, the real-estate transaction may be a real-estate related service.

As used herein, the term “real-estate related service” refers to any type of service and/or assistance that is connected to conducting a real-estate transaction, but which is not the real-estate transaction itself. Non-limiting examples of real-estate related services include legal service, preapproval, insurance related service, renovation services, moving, loan, mortgage, title escrow or any combination thereof. Each possibility is a separate embodiment.

As used herein the term “engage in a real-estate transaction” refers to the readiness of a prospective client for a “next step” such as, being sent real-estate transaction material, engaging in a chat bot conversation, conducting a phone call with an agent, being shown a property, conducting a real-estate transaction, finalizing a real estate transaction and the like.

According to some embodiments, increasing a likelihood that a prospective client will engage in a real-estate transaction comprises increasing the likelihood that the prospective client will ultimately transact. As used herein, the term “ultimately transact” may refer to the client conducting a real-estate transaction within half a year, within 3 months, within 2 months, within 1 month or within 2 weeks or any other timing from the computing of the probability. Each possibility is a separate embodiment.

According to some embodiments, the computer implemented method and/or platform is configured to obtain and/or extract data concerning the client. According to some embodiments, the data may include one or more characteristics of the prospective client. According to some embodiments, the term “obtain” with regards to the one or more characteristics, may refer to passively getting the characteristics, for example from a real-estate agent or directly from the prospective client (e.g. via a questionnaire or through a chat hot). According to some embodiments, the term “retrieve” with regards to the one or more characteristics, may refer to actively recovering, collecting, downloading, saving in folders or otherwise gathering the characteristics, preferably from websites, emails, text messages, video clips or the like, for example, by automatically extracting data, from the internet (e.g., from social media profiles) using application program interfaces (API) and/or natural language processing (NLP) models.

It is understood that at least some of the characteristics, such as various psychological aspects of a prospective client may be hidden/latent, at times even to the subject him/her-self, and may thus be difficult to retrieve directly since they are typically not vocalized. According to some embodiments, the classifying of the prospective client comprises building a psychological profile of the client, based on the obtained and retrieved data, thereby advantageously minimizing the risk of the client making a mistake. A further advantage of building a comprehensive client profile, is that it enables computing a NBA, including non-human interactions, which, due to its personalization, resembles human interaction and even enables creating a trust-relationship, despite its automation.

According to some embodiments, the one or more characteristics may be selected from a life circumstance or change therein, age, gender or gender affiliation, family status, socioeconomic status, employment status, current housing status, geographic location, demographic status, mortgage pre-approval, financial status, financial proof provided, previous transactions or any combination thereof. Each possibility is a separate embodiment.

For example, certain life circumstances (e.g. a newborn child) may, on the one hand, be a strong indicator of a client wanting to transact, yet may, on the other hand, be associated with a low level of engagement/responsiveness to interactions and may require a specific type of nurturing.

As used herein, the term “extracting” with regards to data regarding a digital interaction behavior may refer to collecting, downloading, saving in folders or otherwise gathering digital interaction data. According to some embodiments, the extracting comprises applying an NLP model, an image analysis algorithm, a voice analysis algorithm on the interaction. According to some embodiments, the digital interaction may be passively obtained (e.g. emails with the prospective client obtained from a real estate agent). Additionally or alternatively, the digital interaction may be an Al-mediated interaction, actively initiated with the prospective client, via the platform. According to some embodiments, the data indicative of a digital interaction behavior may be retrieved from pages that can be accessed by a Web browser. According to some embodiments, extracting data from the internet may comprise automatically extracting, using a machine learning model and/or NLP model capable of scrolling and reading the internet.

According to some embodiments, the computer implemented method further comprises identifying any missing information regarding the prospective client and of requesting the prospective client to provide same.

Once the initial gathering of data is completed, the computer implemented method and/or platform may apply one or more machine learning modules on the extracted data to determine an initial real-estate preference associated with the prospective client.

As used herein, the term “initial real-estate preference” refers to a first classification of the prospective client based, for example, on the type of property he/she is pursuing (e.g. apartment or house), the type of transaction (e.g. pre-approval, buying renting), geography, a first budget estimate and the like. According to some embodiments, the initial real-estate preference may further provide an initial indication of a best nurturing strategy.

After an initial real-estate preference has been established, the computer implemented method and platform is further configured to conduct one or more interactions with the prospective client and extracting from the interaction data regarding at least one digital interaction feature of the prospective client indicative of his/her digital interaction behavior and/or preference.

As used herein, the term “digital interaction” may refer to browsing, texting, voice messaging, image messaging, multimedia messaging, bot conversations, comments and/or responses (e.g. to checked property listings), video clips watching, voice calls, video calls or any combination thereof. Each possibility is a separate embodiment.

Advantageously, the interaction may be with a conversation machine which is able to understand complex sentences, some with lots of emotional, suggestive information, and at times even irrelevant information. Moreover, the answers provided by the conversation machine are also complex indicating that the client is “understood” and his preferences/requirements taken into consideration, at times even when not vocalized. According to some embodiments, the conversation machine not only understands what real-estate transaction the person is looking for but makes the best real-estate match according thereto, but also understands the readiness to transact in order to decide what to do next.

As used herein, the term “digital interaction behavior” may refer to qualitative (e.g. content) and/or quantitative (e.g. duration, frequency and speed) of the digital interaction.

As used herein, the term “digital interaction preference” may refer to a preferred type of interaction, for example, whether the prospective client prefers emails, text messages or voice calls.

According to some embodiments, an initial digital interaction may be initiated before the computing/determining of a digital interaction behavior and/or preference of the prospective client. According to some embodiments, the type of the initial interaction may be standardized (e.g. always a text message). According to some embodiments, the initial interaction may be determined based on the first classification of the client.

According to some embodiments, additional interactions may be conducted prior to computing the digital interaction behavior and/or preference of the prospective client.

According to some embodiments, the computer implemented method and platform is further configured to apply the one or more machine learning modules on the initial real-estate preference, and on the at least one digital interaction feature extracted from the one or more interactions, so as to classify the prospective client into a client-type cluster.

Non-limiting examples of features that may be extracted include: an amount of time spent browsing, a degree to which the browsing was an in-depth browsing (e.g. how much time is spent on a particular listing), a quantity of property listings checked, a degree of similarity between the properties checked, a degree of consistency between the property listings checked, preferences investigated by the prospective client, a level of responsiveness by the prospective client, for example, to conversational bots, text messages, multimedia messages, voice messages, voice calls, and/or push notifications, a timing of the prospective client’s response to conversational bots, text/multimedia/voice messages, voice calls and/or push notifications (e.g. in terms of when in the day and/or in terms of also how fast the prospective client interacts after an interaction has been initiated) a listing checking rate, a realistic expectations index, a level of action-taking related to a potential real estate transaction, an amount of people involved by the prospective client with regards to the property listing (e.g. inviting a relative can be positive indication), an amount of sharing of the checked property listing, or any combination thereof. Each possibility is a separate embodiment.

According to some embodiments, the digital interaction features may be derived by applying feature selection and/or feature extraction algorithms on the extracted data. According to some embodiments, deriving the features comprises applying an NLP model on the extracted data. According to some embodiments, deriving the features comprises applying voice analysis, text analysis, voice to text transcription and analysis thereof, sentiment analysis or any combination thereof on the extracted data. Each possibility is a separate embodiment.

Various Al models may be applied on the data to derive features therefrom.

According to some embodiments, speech recognition/transcription models may be applied to audio recordings and/or audio messages to generate a transcribed text. Non-limiting examples of suitable transcription algorithms include AWS transcribe API, Speech-to-Text API, etc. Each possibility is a separate embodiment.

According to some embodiments, NLP models may be applied on text to retrieve specific information from the text, identify keywords or key points in the text and the like. According to some embodiments, the one or more NLP models may include one or more autoregressive language models. According to some embodiments, the one or more NLP may be selected from: Bidirectional Encoder Representations from Transformers (BERT), Robustly Optimized BERT Pretraining Approach (RoBERTa), GPT-3, ALBERT, XLNet, GPT2, StructBERT, Text-to-Text Transfer Transformer (T5), Efficiently Learning an Encoder that Classifies Token Replacements Accurately (ELECTRA), Decoding-enhanced BERT with disentangled attention (DeBERT) Dialog Flow, RASA, Spacy based models, or any combination thereof. Each possibility is a separate embodiment.

According to some embodiments, the classification may be based on real-estate associated preferences.

Non-limiting examples of real-estate associated client types include the ‘parent type’, the ‘party type’, the ‘culturally interested client’, the ‘avantgarde type’, the ‘luxury -type’, the ‘cost sensitive -type’, the ‘social type’, ‘the environmentally aware type’, the ‘privacy concerned type’, the ‘commuting type’, the ‘aggressive type’, the ‘polite type’, ‘urban type’, ‘country type” etc. It is understood by one of ordinary skill in the art that a prospective client may be classified into more than one class.

Additionally or alternatively, the classification may be based on communication preferences. Non-limiting examples of communication associated client types include the “talkative type”, the “reserved type”, the “personal contact type”, the “digital interaction type”, the “text lover”, the “image lover”, the “statistics lover”, the “night-owl type”, the “morning type” etc.

According to some embodiments, the computer implemented method and platform are further configured to compute a customer relationship management (CRM) scheme for the prospective client, based on the client type cluster, into which the prospective client is classified.

As used herein, the term “CRM scheme” refers to streamlined processes according to which interactions with customers are administered. According to some embodiments, the CRM scheme comprises at least one interaction with the prospective client, wherein a time, a type and a content of the subsequent interaction is determined, based on the client type cluster.

As used herein, the term “time” with regards to the interaction may refer to a timing of the interaction including time of the day (e.g. evening or morning), a frequency (e.g. every week, every day), duration (e.g. an up to 2 minute’s interaction). According to some embodiments, the computer implemented method and platform may be configured to calculate a point in time when a specific likelihood to engage in the transaction is reached and to adjust the CRM scheme accordingly. According to some embodiments, the computer implemented method and/or platform is configured to predict a best timing of a next interaction-initiation, based on the derived one or more digital interaction features and the clustering.

As used herein, the term “type” with regards to the interaction may include whether the interaction is automated or human. According to some embodiments, suitable types of interactions include: texting, voice messaging, image messaging, multimedia messaging, bot conversations, voice calls, video calls or any combination thereof. Each possibility is a separate embodiment. According to some embodiments, the computer implemented method and/or platform is configured to predict a best type of a next interaction-initiation, based on the derived one or more digital interaction features and the clustering. As used herein, the term “content” with regards to the interaction may refer to the type of listings presented to the prospective client, the type of images presented to the prospective client, the amount of text included in a written interaction, the number of images presented during an interaction, the language (e.g. English, Spanish etc.), language level (e.g. formal versus informal), or any combination thereof. Each possibility is a separate embodiment. According to some embodiments, the computer implemented method and/or platform is configured to predict a best content of a next interaction-initiation, based on the derived one or more digital interaction features and the clustering.

According to some embodiments, computing the CRM scheme may further include matching a specific agency and/or to a specific agent (or agent type) to the prospective client, based on the classification of the prospective client. Advantageously matching the right agent to the client may increase the probability of transaction and/or reduce time to transaction. According to some embodiments, the output may include both a general and an agent specific probability.

According to some embodiments, the computer implemented method and platform are further configured to adjust the CRM scheme based on additional data extracted from the ongoing interaction with the prospective clients.

According to some embodiments, the computer implemented method and/or platform is configured to predict the engagement of the prospective client in a real-estate transaction based on the timing, type and content of an interaction. For example, the likelihood of engagement may be larger for an in-person telephone call as compared to a chat bot conversation. According to some embodiments, computing the CRM scheme may further include taking into consideration the prediction. For example, if a client is determined by the machine learning to be very likely to engage in a real-estate interaction as a result of a phone call, but has almost no likelihood of engaging if the interaction is in the form of a chat bot conversation, the CRM scheme may be computed accordingly. However, if only a small difference in likelihood of engagement is observed, at least a portion of the interactions in the CRM scheme may be chat bot conversation, thereby significantly reducing the overall human workload.

According to some embodiments, the computer implemented method and/or platform is configured to apply a machine learning module capable of computing a likelihood of transaction. According to some embodiments, the computed likelihood of transaction may be included as an input for the machine learning algorithm configured to classify the prospective client into a client type cluster and/or as an input for the computing of the CRM scheme.

According to some embodiments, determining a likelihood that a prospective client will engage in a real-estate transaction comprises determining the likelihood over time (also referred to as a time-frame likelihood). As a non-limiting example, the computer implemented method may compute that a time-frame probability of engagement of 50% is likely to happen 15 days from an interaction, and a time-frame probability of engagement of 85% in 37 days from the interaction. According to some embodiments, based upon such calculations, time -points of interaction may be automatically set. According to some embodiments, the time-frame probability of engagement may be computed for a multiplicity of scenarios. As a non-limiting example, the computer implemented method may compute that a time-frame probability of engagement of 50% is likely to happen 12 days from an interaction, but that the probability of engagement within that timeframe will rise to 65% if a message is sent 7 days after the interaction. As another non-limiting example, the computer implemented method may compute that a time-frame probability of engagement of 50% is likely to happen 15 days from an interaction, and that the time-frame probability of engagement will rise to 80% in 40 days from the interaction, provided a pushnotification is sent in the intervening time -period. According to some embodiments, a probability may be computed for more than one transaction type. For example, a first probability may be computed regarding the likelihood that a prospective client will engage in a transaction that involves the buying of a property and a second probability regarding the likelihood that the prospective client will engage in a transaction that involves renting a property.

According to some embodiments, the computer implemented method and/or platform is configured to determine one or more unique preferences of the prospective client. This may enable adjusting a client-type specific CRM scheme (e.g. the time, the type and/or the content of subsequent interactions) based on the specific preferences of a specific prospective client. As a non-limiting example, if a prospective client is classified by the machine learning algorithm as being a luxury-type client who favors Instagram messages with lots of pictures, and in addition as having an interest in luxury cars, the interaction of the client may be adjusted to include images of luxury cars at the real-estate listings sent/presented to him/her. According to some embodiments, the computer implemented method and/or platform may be further configured to relax one or more of the unique real-estate preferences, when the preferences result in a low inventory. For example, a client classified as a cost sensitive urban type client with a unique preference of nearby, (this sentence needs correction as it appears to be missing something) the having a unique preference of the property including a bicycle storage room, the bicycle storage room may be amended to common storage facility or entirely removed when none or few properties fitting the criteria are retrieved.

According to some embodiments, the computer implemented method and/or platform may be further configured to identify one or more blockers to transaction, based on the client type cluster and on the one or more unique preferences. Non-limiting examples of blockers to transaction include, difficulty of knowing the budget available, lack of transportation means, fear of neighbors or the like. According to some embodiments, the computer implemented method and/or platform may be further be configured to retrieve a resolution to the identified one or more blockers. As a non-limiting example, budget calculation may be proposed to a client having difficulty of calculating his/her budget. As another non-limiting example, nearby social events may be retrieved by the algorithm and proposed to a prospective client worried about the social network around a certain property.

According to some embodiments, the one or more machine learning models may be a machine learning algorithm trained on a training dataset comprising digital interaction features and user characteristics of a population of clients as well as labels associated with each member of the population of clients, the labels indicating the type of interaction conducted with the client and whether or not the client engaged in the real-estate transaction. According to some embodiments, the machine learning algorithm may be subsequently validated on a validation dataset comprising digital interaction features and user characteristics of a second population of clients.

Reference is now made to FIG. 1, which schematically shows an outline of the herein disclosed Al-based platform for determining a likelihood of a prospective client to conduct a real estate transaction, according to some embodiments.

Starting at the top right side of the figure, a prospective client or, more preferably, a plurality of leads regarding prospective clients (such as 10, 20, 50, 100 or more leads) may be obtained and/or retrieved. As stated herein, the leads may, for example, be obtained from a real- estate agent and/or retrieved from a dedicated real-estate website. The leads serve as an input to the platform.

The platform may then receive and/or retrieve data regarding the prospective client including characteristics, such as age, employment, family status, etc. It is understood that at least a portion of the information may be inputted with the inputted lead. Additionally or alternatively, at least a portion of the information may be inputted by the prospective client him/herself, for example, in response to a questionnaire, a chat bot conversation and/or the like. Additionally or alternatively, at least a portion of the information may be retrieved computationally, for example, from social media and/or other online information.

Then, one or more machine learning modules may be applied on the obtained/retrieved data to determine an initial real-estate preference associated with the prospective client. As for the characteristics, as least a portion of the data may be received along with the lead (for example, from a real-estate agent or agency), may be inputted by the prospective client him/herself, for example, in response to a questionnaire, a chat bot conversation and/or retrieved computationally, for example, from social media and/or other online information.

Based on the initial real-estate preference one or more interactions are then conducted with the prospective client preferably via a conversation machine, as essentially disclosed herein. From the interaction, data regarding at least one digital interaction feature of the prospective client is extracted, preferably by using a dedicated NLP model. Preferably more than one and, optionally, a plurality of features (e.g. at least 4, 5, 10 or more features) are extracted. According to some embodiments, the features may be extracted from voice calls by applying voice analysis and/or speech recognition algorithms thereon. According to some embodiments, the features may be extracted from text (e.g. emails or text messages) by applying NLP models thereon. According to some embodiments, the features may be extracted from video communication (e.g. video calls) by applying image analysis algorithms thereon.

One or more machine learning modules are then applied on the initial real-estate preference, and on the at least one extracted digital interaction feature, to classify the prospective client into a client type cluster and to compute a customer relationship management (CRM) scheme for the prospective client, wherein the CRM scheme comprises at least one subsequent interaction with the prospective client, the time, type and content of which is at least partially determined, based on the client type cluster into which the prospective client is classified.

Reference is now made to FIG. 2A, which is an exemplary flowchart of a computer implemented method 100 for personalized nurturing of prospective real-estate clients, according to some embodiments.

Method 100 may include a step 110 in which the processing unit obtains and/or retrieves a prospective client or more preferably a plurality of leads regarding prospective clients (such as 10, 20, 50, 100 or more leads). As stated herein, the leads may, for example, be obtained from a realestate agent and/or retrieved from a dedicated real-estate website. The leads serve as an input to the platform.

In step 120, data regarding the prospective client(s) may then be obtained or retrieved. It is understood that at least a portion of the data, such as certain client characteristics, may be received along with the lead. Additionally or alternatively, at least a portion of the information may be provided by the prospective client him/herself, for example, in response to an online questionnaire, a chat bot conversation and/or the like. Additionally or alternatively, at least a portion of the information may be retrieved computationally, for example, from social media and/or other online information.

In step 130 one or more trained machine learning models and/or big data analysis models is applied on the extracted data to determine an initial real-estate preference associated with the prospective client, such as, but not limited to, the type of property he/she is pursuing, the type of transaction, the desired geography of the real estate, a desired budget frame and the like. Each possibility is a separate embodiment. Non-limiting examples of suitable big-data analytics include but are not limited to linear regression, Logistic Regression, Classification and Regression Trees, K-Nearest Neighbors, K-Means Clustering, fuzzy models and the like. Each possibility and combination of possibilities is a separate embodiment.

Non-limiting examples of suitable algorithms include convolutional neural network (CNN), recurrent neural network (RNN), long-short term memory (LSTM), auto-encoder (AE), generative adversarial network (GAN), Reinforcement-Learning (RL) and the like, as further detailed below. In other embodiments, the specific algorithms may be implemented using machine learning methods, such as support vector machine (SVM), decision tree (DT), random forest (RF), and the like. Each possibility and combination of possibilities is a separate embodiment. Both “supervised” and “unsupervised” methods may be implemented.

In step 140 one or more initial interactions is conducted with the prospective client. According to some embodiments, the time, type and/or content of the initial one or more interactions may be standardized. According to some embodiments, the time, type and/or content of the initial one or more interactions may be determined based on the determined initial real-estate preference.

In step 150 data regarding at least one digital interaction feature of the prospective client is extracted from the interaction, for example by using various NLP models and/or image/voice analysis algorithms; and in step 160 one or more machine learning models is applied on the extracted features in order to classify the one or more prospective clients into one or more client type clusters/groups/classes. Based on the classification, a CRM scheme personalized to the one or more prospective clients is computed (step 170). According to some embodiments, the CRM scheme may be computed, e.g. in terms of timing, type and content based only on the classification. According to some embodiments, the CRM scheme may be computed based on the client type classification and then further adapted, based on preferences that are unique to a certain prospective client, as essentially described herein.

According to some embodiments, the machine learning algorithm may be trained on an unsupervised dataset comprising a plurality of prospective clients and their associated data, classifications and/or interactions. The unsupervised machine learning discovers patterns and may cluster the data. According to some embodiments, the machine learning algorithm may be trained on a supervised dataset comprising a plurality of digital interaction behavior features and client characteristics and interaction data.

According to some embodiments, the output may be in the form of a document, including, for example, a list of classified prospective clients and their associated CRM scheme (including, for example, a plurality of CRM tasks such as future interactions to be conducted. According to some embodiments, the output may be user-modifiable, e.g. enable a user to filter, sort or otherwise present the output according to his/her preferences.

Reference is now made to FIG. 2B, which is an exemplary flowchart of a computer implemented method 200 for personalized nurturing of prospective real-estate clients including a closed-loop feedback, according to some embodiments. Steps 210-270 are essentially similar to steps 110-170, respectively. In addition, method 200 includes an additional step 280 at which an Al-based digital follow-up interaction (such as a chat bot conversation, a text message or the like) is conducted with the one or more prospective clients. Features are then derived from the followup interaction and the probability of transaction recalculated by returning to steps 260 and 270. It is understood that steps 260-280 may be conducted numerous times until the client is classified as being mature for transaction, in which case the client details are preferably transfered to a real estate agent in an automated manner. According to some embodiments, the real estate agent is the agent who inputted the lead into the herein disclosed computer implmented method. According to some embodiments, a real estate agent is matched to each of the one or more prospective clients, based on the CRM scheme and the client type classification of the client.

Reference is now made to FIG. 3, which is an exemplary flowchart of a computer implemented method 300 for personalized nurturing of prospective real-estate clients (leads), according to some embodiments. In step 310, basic details, such as name and address of a prospective client are obtained. While method 300 is illustrated for a single client, it is understood that it can be carried out on a large plurality of clients (e.g. at least 100 clients, at least 500 clients, at least 1,000 clients, at least 5,000 clients, at least 10,000, at least 50,000 or at least 100,000 clients) simultaneously or in tandem. Each possibility is a separate embodiment.

According to some embodiments, the method is performed on at least 100,000 leads per day, at least 500,000 leads per day or at least 1 ,000,000 leads per day. Each possibility is a separate embodiment.

Additional data is then, in step 320, obtained/extracted or retrieved. According to some embodiments, the additional data may include one or more characteristics of the prospective client. Non-limiting examples of additional data that may be obtained/extracted or retrieved include: life circumstance or change therein, age, gender or gender affiliation, family status, socioeconomic status, employment status, current housing status, geographic location, demographic status, mortgage pre-approval, financial status, financial proof provided, previous transactions or any combination thereof. Each possibility is a separate embodiment. According to some embodiments, at least some of the characteristics may be hidden/latent and as such are not vocalized or otherwise expressed by the prospective client. According to some embodiments, the at least some of the additional data is extracted from pages that can be accessed by a Web browser. According to some embodiments, the extracting comprises applying one or more NLP algorithms.

In step 330, one or more machine learning (ML) models may be applied on the extracted data to determine an initial real-estate preference associated with the client.

Non-limiting examples of suitable algorithms include convolutional neural network (CNN), recurrent neural network (RNN), long-short term memory (LSTM), auto-encoder (AE), generative adversarial network (GAN), Reinforcement-Learning (RL) and the like, as further detailed below. In other embodiments, the specific algorithms may be implemented using machine learning methods, such as support vector machine (SVM), decision tree (DT), random forest (RF), and the like. Each possibility and combination of possibilities is a separate embodiment. Both “supervised” and “unsupervised” methods may be implemented. According to some embodiments, the ML model may be trained on data and/or updated on data extracted from more than 1000, more than 5000, more than 10,000, more than 50,000, more than 100,000 or more than 500,000 prospective clients. Each possibility is a separate embodiment.

Then in step 340 one or more digital interactions may be conducted with the prospective clients. Additionally or alternatively, transcripts of one or more digital interactions previously conducted with the prospective clients may be obtained. Client. From the one or more digital interactions data regarding at least one digital interaction feature of the client may be extracted as well as a feedback of the prospective client to the digital interaction. According to some embodiments, the at least one digital interaction feature is selected from, an amount of time spent browsing per property, a quantity of property listings checked, a degree of similarity between the properties checked, a degree of consistency between the property listings checked, an amount of sharing of the checked property listing or any combination thereof. Non-limiting examples of feedbacks that can be extracted from the digital interaction include: a response of the prospective , client to a content of the digital interaction (e.g., by applying a text analysis algorithm thereon), a sentiment of the prospective client with respect to the content of the digital interaction (e.g., by applying sentiment analysis on the text and/or video of the digital interaction), a request of the prospective client with regards to future digital interaction content and any combination thereof (e.g., based on a response to a follow up questionnaire). In step 350, the initial real-estate preference, the at least one digital interaction feature and the user feedback may be further processed by applying one or more machine learning (ML) models and/or big data analytics thereon. According to some embodiments, the processing comprises computing a likelihood of transaction for the client.

Non-limiting examples of suitable big-data analytics include but are not limited to linear regression, Logistic Regression, Classification and Regression Trees, K-Nearest Neighbors, K- Means Clustering, fuzzy models and the like. Each possibility and combination of possibilities is a separate embodiment.

Moreover, in step 360, a next best action (NBA) for the prospective client is computed based on the processed data. According to some embodiments, the NBA is configured to increase the likelihood of transaction of the prospective client, as determined from the applying of the one or more ML models. Non-limiting examples of suitable NBAs include: asking for more information or clarification on preferences of the prospective client, suggesting an open house, notifying the prospective client on changes in market, asking an advisor to contact the client and any combination thereof. Each possibility is a separate embodiment. According to some embodiments, a time, a type and a content of the NBA is outputted by the one or more ML models, based on the processing of the initial real-estate preference, digital interaction feature and user feedback. According to some embodiments, the NBA may be a digital interaction in the form of: texting, voice messaging, image messaging, multimedia messaging, bot conversations or any combination thereof. According to some embodiments, determining the content of the NBA comprises selecting format, image quality, language, language level, or any combination thereof of the NBA. Each possibility is a separate embodiment. According to some embodiments, determining the timing of the NBA comprises determining a suitable period from previous interaction, time of day, time of week and the like. Each possibility and combination of possibilities is a separate embodiment.

Non-limiting examples of suitable NLP models include: Bidirectional Encoder Representations from Transformers (BERT), Robustly Optimized BERT Pretraining Approach (RoBERTa), GPT-3, ALBERT, XLNet, GPT2, StructBERT, Text-to-Text Transfer Transformer (T5), Efficiently Learning an Encoder that Classifies Token Replacements Accurately (ELECTRA), Decoding-enhanced BERT with disentangled attention (DeBERT) Dialog Flow, RASA, Spacy based models, or any combination thereof. Each possibility is a separate embodiment.

In step 370, the NBA in the form of an additional digital interaction may be automatically executed, so as to increase the likelihood of transaction of the prospective client (nurturing of the client).

It is understood that steps of method 300 may be repeated. For example, additional digital interaction features may be extracted, from the executed NBA, the one or more ML models reapplied, in order to compute a new likelihood of transaction, which in turn may serve as an input for further nurturing of the client. According to some embodiments, the one or more ML models may be continuously updated both in a general manner (based on data obtained from a large plurality of clients) or specifically with regards to a specific client based on a plurality of interactions (personalization of the ML model).

According to some embodiments, steps of method 300 may be repeated until a specific/desired likelihood to engage in the transaction is reached. According to some embodiments, once the desired likelihood of transaction is reached, the client details may (preferably automatically) passed over to a real-estate agent. According to some embodiments, passing over the details comprises matching an agency and/or an agent (type) to the client, determined by the one or more ML models.

According to some embodiments, the nurturing processes is iterative and can run for weeks and even months until the lead is ready to close or almost ready to close a deal (i.e. ready for transfer to a real-estate agent). In the description and claims of the application, the words “include” and “have”, and forms thereof, are not limited to members in a list with which the words may be associated.

It is appreciated that certain features of the disclosure, which are, for clarity, described in the context of separate embodiments, may also be provided in combination in a single embodiment. Conversely, various features of the disclosure, which are, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable subcombination or as suitable in any other described embodiment of the disclosure. No feature described in the context of an embodiment is to be considered an essential feature of that embodiment, unless explicitly specified as such.

Although stages of methods according to some embodiments may be described in a specific sequence, methods of the disclosure may include some or all of the described stages carried out in a different order. A method of the disclosure may include a few of the stages described or all of the stages described. No particular stage in a disclosed method is to be considered an essential stage of that method, unless explicitly specified as such.

Although the disclosure is described in conjunction with specific embodiments thereof, it is evident that numerous alternatives, modifications and variations that are apparent to those skilled in the art may exist. Accordingly, the disclosure embraces all such alternatives, modifications and variations that fall within the scope of the appended claims. It is to be understood that the disclosure is not necessarily limited in its application to the details of construction and the arrangement of the components and/or methods set forth herein. Other embodiments may be practiced, and an embodiment may be carried out in various ways.

The present invention may be a system, a method, and/or a computer program product. The computer program product may include a computer readable storage medium (or media) having computer readable program instructions thereon for causing a processor to carry out aspects of the present invention. The computer readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. The computer readable storage medium may be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non-exhaustive list of more specific examples of the computer readable storage medium includes the following: a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a mechanically encoded device having instructions recorded thereon, and any suitable combination of the foregoing. A computer readable storage medium, as used herein, is not to be construed as being transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission media (e.g., light pulses passing through a fiber-optic cable), or electrical signals transmitted through a wire. Rather, the computer readable storage medium is a non-transient (i.e., not- volatile) medium.

Computer readable program instructions described herein can be downloaded to respective computing/processing devices from a computer readable storage medium or to an external computer or external storage device via a network, for example, the Internet, a local area network, a wide area network and/or a wireless network. The network may comprise copper transmission cables, optical transmission fibers, wireless transmission, routers, firewalls, switches, gateway computers and/or edge servers. A network adapter card or network interface in each computing/processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium within the respective computing/processing device.

Computer readable program instructions for carrying out operations of the present invention may be assembler instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state-setting data, or either source code or object code written in any combination of one or more programming languages, for example, JavaScript, Smalltalk, C, C++, TypeScript, Python and R. The computer readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer (or cloud) may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet Service Provider) including wired or wireless connection (such as, for example, Wi-Fi, BT, mobile, and the like). In some embodiments, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate arrays (FPGA), or programmable logic arrays (PLA) may execute the computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present invention.

Aspects of the present invention are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer readable program instructions.

These computer readable program instructions may be provided to a processor of a general- purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. These computer readable program instructions may also be stored in a computer readable storage medium that can direct a computer, a programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer readable storage medium having instructions stored therein comprises an article of manufacture including instructions which implement aspects of the function/act specified in the flowchart and/or block diagram block or blocks.

The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other device to produce a computer implemented process, such that the instructions which execute on the computer, other programmable apparatus, or other device implement the functions/acts specified in the flowchart and/or block diagram block or blocks.

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware -based systems that perform the specified functions or acts or carry out combinations of special purpose hardware and computer instructions.

The descriptions of the various embodiments of the present invention have been presented for purposes of illustration but are not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein was chosen to best explain the principles of the embodiments, the practical application or technical improvement over technologies found in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.

The following examples are presented in order to more fully illustrate some embodiments of the invention. They should in no way be construed, however, as limiting the broad scope of the invention. One skilled in the art can readily devise many variations and modifications of the principles disclosed herein without departing from the scope of the invention. EXAMPLES

Example 1 - client nurturing

The hereindisclosed method was applied on 6,500 “dead” homebuyer leads (at least two months old leads) obtained from Metropolitan Brokers, who searched for a solution to engage the dead leads in a fast, at scale and cost-efficient manner.

As seen from FIG. 4, the herein disclosed method and system (“Localize”) was able to activate 2308 out of the 6500 leads and out of these 37% were classified as engaged buyers (having an initial likelihood of transaction classified as “active”). These 2308 leads were then nurtured using the herein disclosed method and system comprising a plurality of digital interactions, conducted according to the computed NBA, as disclosed herein. This resulted in 860 (37%) having their likelihood of transaction increased to “engaged buyers” which were then transferred to a realestate agent resulting in 6.15% of those closing a deal. Importantly, this rate of deal-closing is 12 times higher than the industry standard of 0.33% matched deals turning into closed deals. Accordingly, these results emphasize the ability of the herein disclosed system and method to a) turn dead leads into active leads; and b) nurture the leads into leads with a high likelihood of transaction.

While certain embodiments of the invention have been illustrated and described, it will be clear that the invention is not limited to the embodiments described herein. Numerous modifications, changes, variations, substitutions and equivalents will be apparent to those skilled in the art without departing from the spirit and scope of the present invention as described by the claims, which follow.

Claims

1. A computer implemented method for personalized nurturing of prospective real -estate clients, the method comprising: a) obtaining a prospective real-estate client; b) obtaining and/or extracting data concerning the client; c) applying one or more machine learning modules on the extracted data to determine an initial real-estate preference associated with the client; d) conducting one or more digital interactions with the client, and extracting from the interaction: data regarding at least one digital interaction feature of the client; wherein the at least one digital interaction feature is selected from, an amount of time spent browsing per property, a quantity of property listings checked, a degree of similarity between the properties checked, a degree of consistency between the property listings checked, an amount of sharing of the checked property listing or any combination thereof; and a user feedback (text/voice) selected from a response of the prospective client to a content of the digital interaction, a sentiment of the prospective client with respect to the content of the digital interaction, a request of the prospective client with regards to future digital interaction content and any combination thereof, wherein extracting the user feedback comprises applying one or more NLP models on the digital interaction, e) processing the extracted initial real-estate preference, the at least one digital interaction feature and the user feedback, by applying one or more machine learning modules thereon; wherein the processing comprises computing a likelihood of transaction for the client based on the processed initial real-estate preference, on the at least one digital interaction feature extracted from the one or more interactions; f) computing a next best action (NBA) for the prospective client, based on the processing, wherein the NBA is configured to increase the likelihood of transaction of the prospective client, wherein the NBA is selected from: asking for more information or clarification on preferences of the prospective client, suggesting an open house, notifying the prospective client on changes in market, asking an advisor to contact the client and any combination thereof, wherein a time, a type and a content of the NBA is determined based on the processed initial real-estate preference, digital interaction feature and user feedback; g) automatically executing the NBA, thereby increasing the likelihood of transaction of the prospective client, wherein the NBA is a machine -based interaction.

2. The method of claim 1, comprising repeating steps d-g until the likelihood to transact exceeds a predetermined threshold value.

3. The method of claim 1, further comprising extracting, from the executed NBA, additional digital interaction features, and reapplying a second machine learning module on the additional digital interaction features, thereby refining the computed likelihood of transaction.

4. The method of claim 1, wherein the extracting data concerning the client comprises automatically extracting from pages that can be accessed by a Web browser.

5. The method of claim 1 , further comprising calculating a point in time when a specific likelihood to engage in the transaction is reached.

6. The method of claim 1, further comprising matching an agency and/or an agent to the client, based on the client type cluster into which the client is classified.

7. The method of claim 1, wherein the type of the NBA is selected from: texting, voice messaging, image messaging, multimedia messaging, bot conversations or any combination thereof.

8. The method of claim 1, wherein the content of the interaction comprises listings, image quality, language, language level, or any combination thereof.

9. The method of claim 1 , further comprising determining one or more unique preferences of the client.

10. The method of claim 9, wherein the time, the type and the content of the NBA is further determined based on the one or more unique preferences.

11. The method of claim 10, further comprising relaxing the real-estate preference and/or the one or more unique preferences, when resulting in a low inventory.

12. The method of claim 1 , further comprising identifying one or more blockers to transaction, based on the client type cluster and on the one or more unique preferences.

13. The method of claim 12, wherein the NBA comprises a resolution to the identified one or more blockers.

14. The method of claim 1, wherein the real-estate preference comprises one or more of: buying/selling of a property, renting/leasing a property and/or financing a property, acquiring a real-estate related legal service, a realestate related insurance related service, renovation service, moving service, a real-estate related financial service or any combination thereof.

15. The method of claim 1, wherein the first and second machine learning modules comprise one or more of a voice analysis model, text analysis model, a voice to text transcription model, a transcript analysis model, a sentiment analysis model, an image analysis model or any combination thereof.

16. The method of claim 1, wherein the client is a subject visiting, entering and/or communicating with a virtual real estate platform.

17. A system for personalized nurturing of a prospective real-estate client, the system comprising a processor configured to: obtain and/or extract data concerning the prospective client; apply one or more machine learning modules on the extracted data to determine an initial realestate preference associated with the prospective client; extract from one or more interactions conducted with the prospective client: data regarding at least one digital interaction feature of the prospective client; wherein the at least one digital interaction feature is selected from an amount of time spent browsing per property, a quantity of property listings checked, a degree of similarity between the properties checked, a degree of consistency between the property listings checked, an amount of sharing of the checked property listing or any combination thereof; and a user feedback (text/voice) selected from a response of the prospective client to a content of the digital interaction, a sentiment of the prospective client with respect to the content of the digital interaction, a request of the prospective client with regards to future digital interaction content and any combination thereof, wherein extracting the user feedback comprises applying one or more NLP models on the digital interaction; process the extracted initial real-estate preference, the at least one digital interaction feature and the user feedback, by applying one or more machine learning modules thereon; wherein the processing comprises computing a likelihood of transaction for the client based on the processed initial real-estate preference, on the at least one digital interaction feature extracted from the one or more interactions; compute a next best action (NBA) for the prospective client, based on the processing, wherein the NBA is configured to increase the likelihood of transaction of the prospective client, wherein the NBA is selected from asking for more information or clarification on preferences of the prospective client, suggesting an open house, notifying the prospective client on changes in market, asking an advisor to contact the client and any combination thereof, wherein a time, a type and a content of the NBA is determined based on the processed initial real-estate preference, digital interaction feature and user feedback; automatically execute the NBA, thereby increasing the likelihood of transaction of the prospective client, wherein the NBA is a machine-based interaction.