FR3046283A1 - AUTOMATIC METHOD AND DEVICE FOR DETERMINING A CLIENT ROUTE IN A MULTICANAL COMMUNICATION SYSTEM - Google Patents

Abstract

The method for determining a customer journey comprises: reception (E10) from an application, notification of an interaction via a communication channel of a client (C) with the communication system, including customer identification information and textual content characteristic of the interaction; the consultation (E30, E180) of a first database (6) and the obtaining (E40, E190) of a customer journey model, this first base comprising pre-established customer journey models each representing a sequence of steps, at least one step being associated with semantic rules; the semantic analysis (E60, E220) of the textual content with the semantic rules associated with at least one step of traversing a obtained customer journey model; and if, at the end of the analysis, a step of traversing a model corresponds to the textual content, the creation (E110) or the update (E260) in a second database (7) of at least a customer journey instance (NEW_INST (C), INSTj (C)) associated with the customer.

Description

BACKGROUND OF THE INVENTION The invention relates to the general field of communications.

It relates more particularly to the determination, or in an equivalent way, the reconstitution of a customer journey carried out within a multichannel communication system. By "customer journey" we mean a sequence of process steps (also called "journey stages") or activities that a customer of a company carries out in order to meet a need. No limitation is attached to the nature of this need: purchase of a product, repair or update of a product, access to a service offered by the company, etc.

In practice today, such a customer journey is characterized by communications, that is to say interactions through the exchange of messages between the customer and the company, via different communication channels (media). managed by it: voice interaction, online conversation (also commonly known as "chat"), email exchange (e-mail), interaction via social networks, visit of a website page, shop visit, phone call etc. Because of the variety of communication channels that can be used by the customer to interact with the company, such a customer journey is qualified in the following description of "multichannel customer journey", and the underlying communication system managed by the company of "multichannel communication system".

A problem that arises for the company in such a context is the reconstitution of a multichannel journey of the customer. Indeed, the different stages of the journey occurring via potentially different communication channels, and managed by applications or entities of the company's communication system separate or independent, it is difficult for the latter to know the steps of the journey. through which a customer passes to meet his needs. However, this information is valuable for the company, in particular to provide an appropriate and personalized response to the customer, ensure consistency in the processing that is applied to him via the various communication channels, and size its communication system thanks in particular to a better knowledge of the typical paths of its customers, as well as critical points (eg malfunctions, risks) and decision points associated with these routes.

In the current state of the art, there are algorithms (for example the alpha algorithm) for reconstructing and / or discovering business processes by analyzing activity histories (also commonly referred to as "logs") provided. by applications supporting these processes. These so-called process mining algorithms (or "process mining" in English) are based on the assumption that it is possible to record events sequentially, each event referring to an activity (ie to a specific step of 'a process') and being linked to a particular instance of this process. However, the reconstruction of a process enabled by these algorithms relies on a precise log format and obtaining all log events. It is therefore particularly sensitive to noise.

In addition, it uses the application logs as the sole source of data. But today, as mentioned earlier, the business processes implemented by companies increasingly call on communication tools, particularly in customer / business interactions, based on various and varied communication channels with the customer. This "multichannel" dimension, which most of the time results in unstructured data exchanges between the actors of the process, is not currently taken into account by the "process mining" algorithms of the state of the art.

The document US 9 092 801 describes a method for predicting a customer journey from a customer interaction with a multichannel communication system and predetermined customer journey models, in particular with a view to proactively providing this customer with services identified from the client's predicted needs and according to his interests. However, this document does not propose any solution to determine, in other words to reconstruct, from interactions of the client with the communication system, the journey actually taken by this customer.

OBJECT AND SUMMARY OF THE INVENTION The invention responds in particular to this need by proposing an automatic method for determining a customer journey in a multichannel communication system, this method being intended to be implemented by a computing device and comprising: A step of receiving, from a communication application, a notification of an interaction via a communication channel of a client with an entity of the communication system, this notification comprising at least one piece of information enabling identify the client and textual content that is characteristic of this interaction; A step of consulting a first database and obtaining at least one customer journey model, the first database comprising pre-established customer journey models, each pre-established customer journey model representing a sequence of itineraries; determined path steps, at least one path step being associated with a predetermined set of semantic rules; A step of semantic analysis, for at least one obtained customer journey model, of the textual content with the semantic rules associated with at least one step of the journey of the customer journey model; If at the end of the semantic analysis step, it is determined that at least one step of the journey of a client journey model obtained corresponds to the textual content, a step of creation or updating, in a second database, at least one customer journey instance associated with a customer identifier, this journey instance corresponding to a customer journey in progress by the customer.

Correlatively, the invention also relates to a computing device configured to automatically determine a customer journey in a multichannel communication system, this computing device comprising: a reception module able to receive, from a communication application, a notification of an interaction of a client with an entity of the communication system via a communication channel, this notification comprising at least one piece of information making it possible to identify the client and a textual content characteristic of the interaction; A consultation module, configured to consult a first database and obtain at least one customer journey model, said first database comprising pre-established customer journey models, each predetermined customer journey model representing a sequence of milestones; determined paths, at least one path step being associated with a predetermined set of semantic rules; A semantic analysis module configured to analyze, for at least one obtained customer journey model, the textual content with the semantic rules associated with at least one step of the journey of the obtained customer journey model; and a creation or updating module, activated if the semantic analysis module determines that at least one step of a journey of a customer journey model obtained corresponds to the textual content, this module being configured to create or to put in a second database, at least one customer journey instance associated with a customer identifier and corresponding to a customer journey in progress by the customer.

Preferably, said at least one client journey instance created or updated during the creation or updating step corresponds to the same object as the interaction notified by the communication application. The update of an instance of customer journey includes for example the addition in the customer journey instance associated with the customer identifier of a journey step corresponding to the textual content. The creation includes for example the creation of an instance of customer journey associated with the customer identifier having as a first step, a first step of a journey of a customer journey obtained corresponding to the textual content.

Interaction via a communication channel of a client with an entity of the communication system means here an exchange of messages between the client and the entity regardless of the support of these messages, ie whatever the channel of borrowed communication (email, chat, request for access to a website, etc.). The entity of the communication system may be a hardware or software device such as for example a server, a call center, a software application, but also a natural person who has to interact with the customer during his customer journey. The invention thus provides the opportunity for a company to reconstruct the path of a customer by updating automatically and almost in real time, instances of course corresponding to this client based on its interactions with the multichannel communication system of the customer. regardless of the channel taken during these interactions.

Advantageously, to manage the multichannel dimension of the customer journey and the potential lack of structured data to qualify and link each of the customer's interactions with the company's communication system, the invention is based on route models. predefined client and on semantic rules associated with the steps of these models. A semantic analysis carried out from these rules on the textual content describing the interaction of the client with the communication system enables the computing device according to the invention to associate the customer's interaction with a customer journey step and thus reconstruct a logical journey of the customer within the communication system of the company. In other words, via the semantic analysis implemented, the invention establishes a link between the content of a communication (interaction) of the client with the company and one or more steps of one or more customer journey.

The customer journey models and the associated semantic rules can be easily defined by a business analyst or more generally by the company based on its knowledge of the process used by the customer. The path models can be represented for example in the form of graphs or flow diagrams whose nodes illustrate either path steps (ie activities) or routing operations (condition, loops, duplication, join, etc.). edges between each node defining a sequencing of the steps of travel between them. All or part of the steps of the journey are also according to the invention semantically annotated by associating semantic rules for use in the analysis of a textual content reflecting a customer interaction with the communication system of the business. It is noted that it is not necessary to associate semantic rules with a customer journey step for which it is possible to have structured data sent back by the communication application which makes it possible to directly associate the interaction of the client notified by the application with this step of the journey.

In a particular embodiment, at least one step of a journey of a customer journey model is associated with at least one semantic concept defined by an ontology. This or these semantic concepts are associated with the semantic analysis module which can then invalidate or affirm that a communication (interaction) between the client and the company realizes the associated journey step.

In known manner, an ontology constitutes a data model representative of one or more semantic concepts as well as the relations existing between these concepts. Such a data model is conventionally used in artificial intelligence and software engineering in particular. It can be represented easily as a graph, where the nodes of the graph represent the semantic concepts and the edges the relations between the concepts. In accordance with this embodiment, the ontology is enriched by semantic rules used to analyze the textual content of an interaction and link it to a customer journey instance.

In a particular embodiment, the determination method further comprises a step of consulting the second database and obtaining at least one path instance associated with a client identifier, the semantic analysis step. textual content being implemented with the semantic rules associated with at least one step of a journey of a customer journey model corresponding to said at least one path instance obtained by consulting the second database. In this embodiment, the invention makes it possible to associate the interaction of the notified client with the computing device with an instance of the customer journey in progress, and to accelerate the semantic analysis by taking into account customer journey models. corresponding to this instance of customer journey. The semantic analysis step can thus notably comprise, for each instance of the journey obtained by consulting the second database, the determination, using said model of customer journey corresponding to this instance of journey and taking into account the last step of course of this instance of course, at least one probable course step following the last step of the course, the textual content being analyzed semantically with the semantic rules associated with said at least one probable course step.

This embodiment makes it possible to accelerate the execution of the semantic analysis.

In a particular embodiment, the method further comprises a step of lemmatizing the textual content characteristic of the interaction preceding the semantic analysis step, the semantic analysis step being implemented on the textual content obtained after lemmatization.

In a known manner, lemmatization refers to a lexical analysis of a textual content that consists of associating together the different forms that a word can take, such as, for example, a name (eg wish), a plural (eg wishes), an infinitive verb (eg, wish), etc. This lemmatization stage aims to transform all or part of the words of the textual content characteristic of the interaction into its "dictionary" version to facilitate its semantic analysis. Thus, in the example given above "I wish to buy two televisions" becomes during this stage "I wish to buy two television". This preliminary step makes it possible to simplify the semantic rules defined for each step of the customer journey, and in particular to reduce the number of rules to be defined.

In a particular embodiment, the notification of the interaction further comprises at least one additional information among: - a temporal information associated with the interaction, - an identifier of the application or the communication entity, - a identifier of an actor who participated in the interaction.

This various information constitutes various information that can be used later by the company during the implementation of its business processes.

In a particular embodiment, the client identifier is obtained from the information contained in the notification to identify the client and from a correspondence table between client identifiers on different channels of the multichannel communication system. .

Indeed, the different channels of the multichannel communication system borrowed by the client may not use a unique identifier for the client. For example, the customer can be identified in these channels by a telephone number, an email address, a web login, an IP address, a cookie, a voice message, an identifier on a social network, etc. The correspondence table makes it possible to establish a link between all these identifiers and to make sure that the customer is well identified to allow the reconstitution of his customer journey.

In a particular embodiment, the various steps of the automatic method for determining a customer journey are determined by computer program instructions.

Accordingly, the invention also relates to a computer program on an information carrier, this program being capable of being implemented in a device for automatically determining a customer journey or more generally in a computer, this program comprising instructions adapted to the implementation of the steps of a method of automatic determination of a customer journey as described above.

This program can use any programming language, and be in the form of source code, object code, or intermediate code between source code and object code, such as in a partially compiled form, or in any other form desirable shape. The invention also relates to a computer-readable information medium or recording medium, comprising instructions of a computer program as mentioned above.

The information or recording medium may be any entity or device capable of storing the program. For example, the medium may comprise storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, or a magnetic recording medium, for example a floppy disk or a disk. hard. On the other hand, the information or recording medium may be a transmissible medium such as an electrical or optical signal, which may be conveyed via an electrical or optical cable, by radio or by other means. The program according to the invention can be downloaded in particular on an Internet type network.

Alternatively, the information or recording medium may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the method in question. The invention also relates to a computer system comprising: a first database comprising customer journey models, each customer journey model representing a sequence of determined journey steps, at least one journey step being associated with a predetermined set semantic rules; A second database comprising client journey instances, each path instance being associated with a client identifier and corresponding to a customer journey in progress by the customer; and a computing device according to the invention able to consult the first database and to create or update at least one instance of the client path in the second database.

It may also be envisaged, in other embodiments, that the automatic determination method, the computing device and the computer system according to the invention present in combination all or part of the aforementioned characteristics.

BRIEF DESCRIPTION OF THE DRAWINGS Other features and advantages of the present invention will emerge from the description given below, with reference to the accompanying drawings which illustrate an embodiment having no limiting character. In the figures: - Figure 1 shows schematically a computer system according to the invention, in a particular embodiment; FIG. 2 illustrates the hardware architecture of a computer device of the system of FIG. 1, according to the invention, in a particular embodiment; FIGS. 3 and 4 show examples of graphical interfaces that can be implemented by the computer device of FIG. 2; and FIG. 5 represents, in the form of a flow chart, the main steps of a method for automatically determining a customer journey, in a particular embodiment in which it is implemented by the computing device whose architecture hardware is illustrated in Figure 2.

Detailed description of the invention

FIG. 1 represents, in its environment, a computer system 1 according to the invention, in a particular embodiment.

According to the invention, the computer system is able to determine (reconstruct) automatically and almost in real time, an instance of customer journey in progress of a customer of a company interacting with it via a system. multichannel communication (not referenced in the figure), to meet a need. By "customer journey" we mean a sequence of process steps (also called "journey stages") or activities that the customer performs to satisfy his needs. These process steps are based on one or more customer interactions with one or more entities (eg server (s), software application (s), company employee (s), etc.) of the system. communication network, via one or more generally distinct communication channels. The communication system of the company thus includes all of the aforementioned entities, which can be both hardware devices or software that physical persons.

As mentioned above, there is no limitation on the nature of the customer's need: purchase of a product, repair or update of a product, access to a service offered by the company, etc.

In addition, no limitation is attached to the communication channels used as support for customer interactions with the company and more particularly with the entities of its multichannel communication system to meet the needs of the customer. These interactions generally take the form of exchanges of one or more messages with the entities of the multichannel communication system by e-mail (e-mail), instant messaging, telephone, voice interaction, access to a website, dialogue with a natural person etc.

In the example envisaged in FIG. 1, a CurrI client path in progress is considered by a client C, this client journey consisting of a plurality of travel steps (three steps in the example illustrated in FIG. 1), each step of the path resulting in an interaction INT1, INT2, INT3 of the client C, via a terminal 2, with the communication system of the company via a communication channel CH.

More particularly, the first step of the CurrI customer journey is performed via an interaction INT1 of the client C, via its terminal 2, with a server 3-1 of the multichannel communication system of the company, this server hosting a software application 4-1 (communication application within the meaning of the invention) configured to interact with the client C to meet its needs. The terminal 2 of the client C and the server 3-1 communicate here via a telecommunications network not shown in the figure.

Similarly, the second step of the CurrI customer journey is performed via an INT2 interaction of the client C via its terminal 2 with a server 3-2 of the company's multichannel communication system, this server hosting a server. software application 4-2 (communication application within the meaning of the invention) configured to interact with the client C to meet its needs. The terminal 2 of the client C and the server 3-2 communicate here via the aforementioned telecommunications network.

Finally, in the example envisaged here, the third step of the CurrI customer journey is performed via an INT3 interaction of the client C via its terminal 2 with a server 3-3 of the company's multichannel communication system. , this server hosting a software application 4-3 (communication application within the meaning of the invention) configured to interact with the client C to meet its needs. The terminal 2 of the client C and the server 3-3 also communicate here via the aforementioned telecommunications network.

For the sake of simplification in the rest of the description, when reference is made to any of the servers 3-1, 3-2 and 3-3, any of the software applications 4-1, 4-2 and 4-3, or any of the INT1, INT2 and INT3 interactions, the general references 3, 4 and INT are used.

No limitation is attached to the nature of the telecommunications network used by the terminal 2 to communicate with the servers 3 and their software applications 4: it may be a mobile or fixed telecommunications network, wireless or wired. Similarly, no limitation is attached to the type of terminal 2 used by the client C to interact with the multichannel communication system of the company. This may be for example a smartphone (or "smartphone" in English), a computer, a digital tablet, etc. Client C can also use separate terminals to interact with the company's communication system during the various stages of its customer journey. Finally, no limitation is attached to the number of steps constituting the CurrI customer journey, and the number of interactions resulting, the number of 3 considered here is given for illustrative purposes.

Each INT interaction between the client C via its terminal 2 and a software application 4 consists here of an exchange of one or more messages between the terminal 2 and the software application 4 via a communication channel CH managed by this application. 4.

In the example envisaged in FIG. 1, the communication channel CH1 for the interaction INT1 for example supports (or includes) the exchange of emails between the terminal 2 and the software application 4-1. The terminal 2 also interacts during the INT2 interaction with the application 4-2 via a communication channel CH2, this channel comprising a telephone call between the terminal 2 and the application 4-2 hosted by the server 3-2, and during the INT3 interaction with the application 4-3 via a communication channel CH3 comprising an access of the terminal 2 to a website hosted by the server 3-3.

Note that the nature of the software application 4 managing a communication channel CH depends on the type of this communication channel CH. For example, for a communication channel CH supporting a voice type telephone call, the application 4 may be a call manager; for a communication channel supporting chat or e-mail type exchanges, the application 4 can be a chat or e-mail application; for a communication channel such as access to a website, the application 4 may be the website in question, etc.

In such a context, an interaction between the terminal 2 and a software application 4 within the meaning of the invention comprises all the messages exchanged during a session established between these two entities. Thus, for example, in the case of an exchange of e-mails, the interaction INT1 corresponds to the first e-mail sent by the terminal 2 to the software application 4 (or vice versa), then to the responses to this first e-mail and to the responses to responses exchanged between these two entities, possibly bounded by a predetermined time window. For a telephone call, the interaction INT2 corresponds to all the voice exchanges between the client C (via its terminal 2) and the operator of the server 3 (which may be a voice application) during this telephone call. For an in-store visit, the interaction consists of all the exchanges between customer C with an employee of the company present in the store, etc. It should be noted that when such interaction between a customer and a natural person is part of the CurrI customer journey, the company's communication system includes a communication application that keeps track of this interaction and its content.

It should be noted that the same client is not necessarily identified by a unique identifier with the servers 3 and the software applications 4 that they host, as well as on the different CH communication channels managed by these applications. On the contrary, it is very likely that the client is associated with an identifier ID1 (eg email address) on the communication channel CH1, an identifier ID2 (eg telephone number) on the communication channel CH2 and an identifier ID3 ( eg IP address) on the CH3 communication channel all separate. No limitation is attached to the nature of these identifiers, which of course depend on the application concerned (email address, phone number, IP address, login, etc.). It is assumed here that these multiple identifiers associated with the same client C are reported in a table of correspondence TAB allowing them to be linked to each other and to the client C.

According to the invention, the computer system 1 is able to link an INT interaction between the terminal 2 and the software application 4 to an instance of customer journey C client being implemented (in this case CurrI course). For this purpose, it comprises: a computer device 5, according to the invention; and - two databases 6 and 7, respectively comprising a set of pre-established customer journey models and a set of customer journey instances associated with various customers of the company.

Note that the databases 6 and 7 can be hosted in the computing device 5 itself.

The term "database" refers to any type of medium that makes it possible to store data in a structured manner. The predefined (ie predefined) customer journey models included in the database 6 can be defined for example, via a graphical interface provided for this purpose, described in more detail later, by a business analyst or by the company, independently of the channel. of communication used. Here they take the form of a flow diagram whose nodes represent either steps of the path (or activities), or routing operations (condition, loop, duplication, join, etc.), whose edges define a sequencing stages of journey between them. Such a model can be defined for example in accordance with the BPMN 2.0 (for Business Process Model and Notation) standard in an XML representation (extensible Markup Language).

For example, for a customer journey related to the purchase of a product or the subscription to a service, the steps of a typical customer journey can be the following ones: - stage STI: the customer seeks to inquire about the product / service and to be contacted or to contact an operator, or step STlbis: an operator of the company contacts the customer about this product / service; the steps STI and STlbis that can be represented in the customer journey model representing this typical customer journey as running alternately (one or the other); - Step ST2: the customer is greeted and informed by the operator of the company, obtains information about the product / service and its operation, and decides to buy the product or subscribe to the service; Step ST3: the customer orders the product or subscribes to the service and makes the corresponding payment if necessary; Step ST4: the customer receives his product or access to the service by this customer is activated; - step ST5: the product / service is installed by the customer, possibly with assistance provided by the company; - step ST6: the product / service is used for the first time by the customer; - step ST7: the customer receives a receipt for the payment of the product / service and proceeds to the payment if this has not already been done.

Various customer journey models are thus defined in the database 6 for different types of processes and services offered by the company (eg purchase of a product or subscription to a service, request for assistance or repair of a service). product, access to a service, etc.), and regardless of the communication channels used to implement the journey steps. These models depend of course on the field in which the company intervenes.

According to the invention, at least one step of a journey of a customer journey model stored in the database 6 is associated with a set of rules of predetermined semantics for qualifying an unstructured textual content and associate it with a step of customer journey. In the embodiment described here, each step of the journey of each customer journey model is associated with one or more concepts defined and linked by an ontology, each concept of the ontology being associated in turn with one or more semantic rules.

However, as mentioned earlier, it is possible that certain steps of customer journey of a customer journey model are not associated with semantic rules. This is particularly the case for the steps for which it is possible, during an interaction between the client and the communication system of the company corresponding to one of these steps, to obtain structured data enabling link the interaction directly to this step. A semantic analysis of the interaction is no longer necessary. For the sake of simplicity in the example envisaged here, it is assumed that each step of the journey of each customer journey model included in the database 6 is associated with a set of semantic rules.

In known manner, an ontology is a data model that can be represented as a graph, the nodes of the graph representing the semantic concepts and the edges the relations between these concepts. According to the invention, this ontology is enriched by the different semantic rules associated with the steps of the journey of the customer journey models.

The definition of the semantic rules associated with each step of the journey can be done for example by the business analyst mentioned above, from previous interaction or feedback analyzes, or by the company directly via an interface. graphic provided for this purpose. In other words, the business analyst annotates through this graphical interface each step of the journey with a set of semantic rules intended to be used to link each interaction of the client C with the multichannel communication system of the company. These semantic rules can be stored in a file.

For example, each semantic rule consists of one or more words intended to be searched for in the interaction between the client and the multichannel communication system in order to link this interaction to a customer journey step, and reconstitute the instance of customer journey in progress of client C, as described in more detail later. For example, it can be found in the form of "search {offer} and {product}" where offer = {propose, suggest, present, hear} and product = {television, tv, tv, home cinema, HC, open, internet offer mobile, origami, phone, mobile, smartphone}. As an illustration, with reference to the above-mentioned example of a customer journey relating to the purchase of a product or the subscription of a service: - a semantic rule associated with the STl / STlbis contact step can for example, to understand the word "offer" in association with the name of the product or service concerned; A semantic rule associated with the step ST2 of information may comprise, for example, the word "inform", another the word "notice", each word being considered in the corresponding rule in association with the name of the product or service concerned; A semantic rule associated with the purchase step ST3 may comprise, for example, the word "buy" associated with the name of the product or service concerned; - etc.

Of course, the customer journey models stored in the database 6 and the semantic rules associated with the journey steps taken by these models can be updated (deletion / addition / modification), in particular via a route model manager 6A. customer. This is indeed configured so that you can perform model backup, read an existing model, modify an existing model and delete a template. The examples given above are also given for illustrative purposes only.

It is also understood that the semantic rules associated with the steps of the journey must preferably be as exhaustive as possible in order to easily identify which step in the path of a customer journey model refers to a customer interaction with the customer service system. company communication.

In the embodiment described here, the computing device 5 has the hardware architecture of a computer, as represented in FIG. 2. It notably comprises a processor 8, a random access memory 9, a read-only memory 10, a non-volatile memory. volatile 11, in which is stored in particular the aforementioned TAB correspondence table establishing a link between the different identifiers of the same customer on the various communication channels used by the company, and a communication module 12 allowing it to communicate in particular with the software applications 4 of the servers 3 of the multichannel communication system of the company.

The read-only memory 10 of the computing device 5 constitutes a recording medium according to the invention, readable by the processor 8 and on which is recorded a computer program PROG according to the invention, comprising instructions for the execution of the steps of a method for automatically determining a customer journey instance according to the invention.

This computer program equivalently defines functional modules (and software here) shown in Figure 1, and configured to implement the steps of the determination method according to the invention. These functional modules support or control the physical hardware elements 8-12 of the computing device 5 described above. They include in particular here: - a receiving module 5A able to receive notifications from the software applications 4 informing the computing device 5 of interactions between the communication system of the company and a customer of the company (customer C here ). The module 5A relies in particular on the communication module 12 shown in FIG. 2; A module 5B for consulting the databases 6 (first database within the meaning of the invention) and 7 (second database within the meaning of the invention), and in particular client journey models and route instances client in progress stored in these databases, as well as semantic rules where appropriate associated with the steps of customer journey models. It should be noted that, alternatively, a consultation module distinct from each of the bases may be envisaged; A semantic analysis module 5C configured to analyze a textual content representative of an interaction between a software application 4 and the client C from semantic rules obtained by the module 5B. This module 5C can rely on known semantic analysis tools, such as for example the semantic analysis tool described in particular in document EP 1 585 110; and a module 5D for creating and updating instances of client traversal in the database 6, activated by the module 5C as a function of the result of the semantic analysis performed by the latter.

In the embodiment described here, the program PROG also defines different graphical interfaces grouped under the notation 5E making it possible, on the one hand, to feed the database 6 with new customer journey models and / or to modify the route models. existing customer, and secondly to consult customer paths, either consolidated, or in real time (instance of customer journey in progress).

Thus, more specifically, in the example illustrated in FIG. 1, the PROG program defines three graphical interfaces: a management interface 5E1, which allows the business analyst or the company to define the route models typical customer and "abstract" (ie theoretical) stored in the database 6, independently of the channel used for the realization of each step of course. This 5E1 interface is programmed here to allow the business analyst or company to define these customer journey models graphically in the form of a sequence of journey steps and annotate each step by concepts. semantics defined in an ontology; A 5E2 interface called consolidated customer consultation consultation, which allows the business analyst or any other authorized user of the company to consult an aggregate view of customer journey instances including the communication channel through which each step of the journey took place. Figure 3 illustrates an exemplary interface 5E2. In this figure, a customer journey implementing steps STI, ST2 and ST3 is represented as well as the various CH communication channels via which these steps have been implemented (eg SOC social networks, WEB website, instant messaging CAT, and TEL phone, other OTH here). This interface can also display quantitative information on customer paths, such as the number of instances that have followed a particular path (i.e. implementation of steps via given communication channels), etc. In the example of Figure 3, 70% of instances have borrowed to implement steps STI, ST2 and ST3 WEB communication channels, SOC NW and WEB respectively, and 30% of instances have borrowed to implement these same steps STI, ST2 and ST3 the communication channels TEL, SOC NW and OTH respectively. A system of zooming on a step, on a channel, on a route of a given client, can also be provided by this interface, as well as information relating to the inter-process sequences (eg x% of service paths trigger a shopping path); A 5E3 interface known as real-time customer consultation: this interface is intended for the operators (eg salesman, call center, etc.) of the company. It makes it possible to display a client file, enriched with the current course instances of the latter. FIG. 4 illustrates an example of such an interface 5E3 representing a customer card of an X client. It includes a space on the left comprising various information on the customer (his name, his various identifiers ID1, ID2, ID3 on the service channels). communication CH1, CH2, CH3 proposed by the company to interact with its communication system, its FCT function within the company, its address ADR, etc.), as well as in a right part, the course instance (s) client in progress by this client. In the example illustrated in FIG. 4, only one instance of path is represented comprising two completed stages STI and ST2, carried out respectively via the communication channels CH1 and CH2, and a probable step ST3. Such a graphical interface advantageously allows an operator of the company in contact with the client X to better know his current needs and to be informed of the information already received via other communication channels by the client X.

The functions of these modules are described in more detail now with reference to FIG.

FIG. 5 represents, in the form of a flow chart, the main steps of a method for automatically determining a customer journey according to the invention, in a particular embodiment in which it is implemented by the computer device 5 of FIG. Figure 1.

In the embodiment described here, in order to speed up the processing of each client C interaction with an entity (eg server 3 or software application 4) of the notified company's communication system with the computing device 5, the method automatic determination according to the invention takes place in two phases PHI and PH2 executed in parallel on the interaction considered. The PHI phase aims to deal more particularly with an interaction which constitutes a first step in the path of a customer journey instance of the customer C and if necessary ends with a creation of a customer journey instance integrating as a first step said interaction. Phase PH2 aims to deal more specifically with other interactions that are not early stages of customer journey, but are linked to instances of customer journey in progress. It therefore ends, if necessary, by updating a customer journey instance associated with the client C. The execution in parallel of these two phases, for example via two processes executed in parallel by the computing device 5, In particular, it makes it possible to more quickly detect the interactions relating to the first steps of instances of the customer journey of the interactions relating to subsequent journey steps.

As a variant, the PHI and PH2 phases can be implemented sequentially.

We will describe the progress of these two phases when they are implemented by the computing device 5 to connect an INT interaction between the terminal 2 of the client C and an entity (server 3 and more particularly software application 4 here) of the system. communication of the company via a communication channel CH, to an instance of customer journey C client in progress. This INT interaction is part of a customer journey step implemented by customer C.

Following the INT interaction between the terminal 2 and the software application 4 of one of the servers 3 of the communication system of the company, the software application 4 sends a notification to the computing device 5 informing it of this interaction . This notification denoted NOTIF (INT) is received by the computing device 5 via its communication module 12 and transmitted to the reception module 5A (step E10).

According to the invention, the NOTIF (INT) notification comprises an INFO_ID information enabling the identification of the client C and a textual content TEXTJNT characteristic of the INT interaction. The information INFO_ID is in the example envisaged here the identifier used by the terminal 2 on the communication channel CH (ID1 for the CH1 channel, ID2 for the ID2 channel or ID3 for the ID3 channel). This information, thanks to the correspondence table TAB stored in the non-volatile memory 11 of the computing device 5, uniquely identifies the client C.

It should be noted that other information to identify the client C that identifiers themselves can be considered alternatively. For example, in the case of a communication channel constituted (or comprising, supporting) a voice interaction between the client C and the application 4, a voice identification and / or facial recognition technique can be implemented. implemented by the computer device 5 from audio and / or video and / or audio-video streams acquired during the INT interaction and transmitted by the application 4 in the notification NOTIF (INT).

The NOTIF (INT) notification may also contain other contextual information relating to the INT interaction, as described later. In particular, it can include the type of communication channel CH used during this interaction.

The textual content TEXT_INT retranscribes in the form of a text the whole of the INT interaction between the client C and the application 4, that is to say, all the messages exchanged between the client C, via its terminal 2 here, and the application 4 via the communication channel CH. For example, in the case of an interaction in the form of an exchange of emails (e-mails) between the client C and the application 4, the content TEXTJNT sequentially repeats the content of all the e-mails exchanged. In the case of an interaction in the form of an oral exchange between the client C and an employee of the company (which does not necessarily involve the terminal 2), the application 4 repeats the textual content of this exchange.

This textual content TEXTJNT is supplied to the semantic analysis module 5C of the computing device 5 for analysis.

In the embodiment described here, this analysis begins with a step of lemmatizing the textual context TEXTJNT (step E20). As mentioned earlier, lemmatization refers to a lexical analysis of textual content that consists in associating together the different forms that a word can take, such as, for example, a name (eg wish), a plural (eg wishes), an infinitive verb (eg, wish), etc. This lemmatisation step E20 carried out by the semantic analysis module 5C thus aims to transform all or part of the TEXTJNT textual content words characteristic of the INT interaction into its "dictionary" version to facilitate its semantic analysis.

Thus, for illustrative purposes, if the textual content TEXTJNT contains the phrase "I wish to buy two televisions", it is transformed during step E20 into "I wish to buy two television" by the module SC, the words in italic showing words modified by lemmatization. The rules intended to be applied during this lemmatisation stage are defined beforehand. It is from the textual content TEXTJNT 'resulting from the transformation by lemmatization TEXTJNT textual content that are performed in parallel the two phases PHI and PH2.

Note that the lemmatisation step is optional and that if it is not implemented phases PHI and PH2 are applied directly to the textual content TEXTJNT.

During the PHI phase, the module 5B for obtaining the computing device 5 firstly consults the database 6 of the customer journey models (step E30). This consultation can be done through the customer journey model manager 6A, via known techniques for querying databases.

The module 5B obtains, following this consultation, one or more models MODl, ..., MODN of customer journey, where N denotes a predetermined integer greater than or equal to 1 (step E40). Each model consists of a sequence of customer journey steps, each step of the customer journey being associated here with semantic rules.

The models MODl, ..., MODN (including the semantic rules associated with the steps of these models) are transmitted to the semantic analysis module 5C.

The semantic analysis module 5C then extracts from each of the models MOD1,..., MODN the input stages of these models (that is to say the first step (s) of each of the models, a model that can have several entry stages, as in the example of customer journey given above for illustrative purposes and related to the purchase of a product or the subscription to a service that includes two stages of entry STI and STlbis).

For the sake of simplicity here, it is assumed that each model stored in the database 6 comprises a single entry step. One denotes by STl_MODi, i = l, .., N, each of the steps of entry thus recovered for each models MODl, ..., MODN (step E50).

For each model MODi, taken sequentially here, the module 5C analyzes the textual content TEXTJNT 'using the semantic rules associated with the first step STl_MODi of this model (step E60).

Note that if the textual content TEXT_INT has not been lemmatized, the 5C module operates directly on this textual content.

Furthermore, if a customer journey model comprises several possible entry steps, the analysis of the textual content TEXTJNT 'is performed independently with respect to each of these entry steps for this model.

If the textual content TEXT_INT 'does not correspond to the semantic rules associated with the first step STl_MODi of the MODi model considered (answer no to the test step E70), the semantic analysis module 5C determines whether there are models that have not yet been considered. or examined, that is to say in respect of which the semantic analysis module 5C has not yet analyzed the textual content TEXTJNT '(test step E80). For example, "does not match" means that the textual content TEXTJNT 'does not contain the word or words defined by the semantic rules considered if they define words to be found in the textual content. In any case, this means that the textual content TEXTJNT 'does not respect the semantic rules considered.

If there is still at least one model not yet examined (answer no to the test step E80), the semantic analysis module 5C chooses a new model from among these unexamined models (step E90 of incrementing the index i) , and resumes the previous analysis steps (E60 and E70) using the semantic rules associated with the first step of this new model.

If the textual content TEXTJNT 'corresponds to the semantic rules associated with the first step STl_MODi of the MODi model considered (yes answer to the test step E70), the semantic module 5C selects this model MODi (step E100). For this purpose, it stores, for example, a reference of the MODi customer journey model (i.e. allowing it to identify or to find the model of customer journey MODi in the database 6) in a set SEL_MOD.

Then it determines if all the customer journey models have been examined (test step E110). If unexamined models remain (answer no to the test step E110), the semantic analysis module 5C chooses a new model from among these unexamined models (step E90 of incrementing the index i), and resumes steps E60 to E110 using the semantic rules associated with the first step of this new model.

If during the test step E80 or during the test step E110, the semantic analysis module 5C determines that all the models have already been examined (yes answer to the test step E80 or to the test step El 10 ), the semantic analysis module 5C determines if models were selected during the step E100, that is to say if the set SEL_MOD contains at least one reference of a selected model (test step E120).

Note that in an alternative embodiment, all models MODi, i = 1, ..., N can be examined in parallel by the semantic analysis module 5C.

If no model has been selected, that is to say here if the set SEL_MOD is empty (yes answer to the test step E120), the phase PHI is completed (step E130).

Otherwise, if at least one model has been selected (answer no to the test step E120), the module 5D of the computing device 5 creates a new customer journey instance NEW_INST (C) associated with the client C from this selected model ( step E140). More particularly, if we denote by iO the index of the model selected, the new instance of customer journey NEW_INST (C) created by the 5D module here contains the first step STl_MODiO model MODiO. A new client path instance associated with the client C is thus created for each selected model identified in the set SEL_MOD. In other words, it is possible for the same customer C to create during step E140 several instances of customer journey, each instance of customer journey being associated with a model (unique) of customer journey.

This new instance of customer journey reflects CurrI customer journey in progress by the customer C. It is stored in the database 7 containing instances of customer journey (step E150). In the embodiment described here, the instance NEW_INST (C) is stored in the database 7 in association with the object CurrI customer journey and with a reference to identify the customer journey model corresponding to this instance . The object of the customer journey can be retrieved by the semantic analysis module 5C in the textual content TEXT-INT using pre-established semantic rules. These rules can coincide with the semantic rules defined for each step of the customer journey (eg the rule then relates to an action and a specific object), or correspond to other semantic rules dedicated to the recognition of the object to which it relates. the customer journey.

In the previously given example of a textual content "I want to buy two televisions", the instance NEW_INST (C) corresponds to the first step of a model of customer journey type purchase, and the object of the customer journey is "Television".

It should be noted that other contextual information can be recorded in association with the client journey instance NEW_INST (C), such as, for example, temporal information (eg temporal information of realization of the INT interaction), an identifier of the application software 4 or server 3 with which the interaction INT took place, and / or an identifier of an actor who participated in this interaction (eg operator of the application 4). The communication channel borrowed during the first step can also be stored. The storage step E150 of the instance NEW_INST (C) in the database 7 closes the phase PHI (step E130).

It is noted that this phase PHI is all the more relevant that the INT interaction notified in step E10 is the first INT1 interaction performed by the client C as part of its CurrI customer journey. It should be noted that if another INT interaction that the INT1 interaction is processed during the PHI phase and this interaction does not correspond to any first step of a customer journey model stored in the database 6, the PHI phase s completes without creating a new instance in database 7.

We will now describe phase PH2.

During the PH2 phase, the obtaining module 5B consults the database 7 of the customer journey instances using the client identifier C obtained from the NOTIF notification (INT) (step E160).

If at the end of this consultation, no customer journey instance associated with the client C is found (answer no to the test step E170), the phase PH2 is completed (step E280).

In the embodiment described here, it is limited to a search in the database 7 of instances of course corresponding to the same object as CurrI customer journey. This object can be obtained as previously described for the PHI phase by semantically analyzing the TEXTJNT 'content using appropriate semantic rules.

If at least one instance of client journey INST (C) associated with client C is obtained (yes answer to test step E170), module 5B queries database 6 to obtain the customer journey model corresponding to this instance of INST (C) path (step E180). It uses for this purpose the associated customer journey model reference in the database 7 Inst instance (C).

Module 5B repeats this operation for each INST (C) client path instance found for client C.

We denote by INSTj (C), j = l, ..., M where M denotes an integer greater than or equal to 1, the instances of customer journey found for client C, and MOD (INSTj (C)), the model obtained customer journey corresponding to each instance Instj (C) (step E190). From the models thus obtained and from the last LASTST travel step stored for each instance INSTj (C), the module 5B determines a next step of probable course of the instance of course INSTj (C) (step E200). More particularly, it determines for each instance INSTj (C), with regard to the model MOD (INSTj (C)) corresponding to this instance, such a probable path step designated by PROBSTj. The module 5B thus obtains at the end of the step E200, M probable path steps noted PROBST1, ... PROBSTM for the M instances INSTj (C), j = 1, ... M respectively. These probable course steps correspond to the steps of course that immediately follow the LASTST course step in each of the models of course MOD (INSTj (C)), j = l, ..., M.

Then the module 5B retrieves, by consulting again the database 6, the semantic rules associated with each of these PROBSTl path steps, ... PROBSTM in the MOD MOD (INSTJ (C)), j = l , ..., M respectively (step E210).

The PROBSTl, ... PROBSTM (including the semantic rules associated with these steps) are transmitted to the semantic analysis module 5C by the module 5B.

For each instance INSTj (C) and model MOD (INSTj (C)) corresponding to this instance, j = l, ... M, taken sequentially here, the module 5C analyzes the textual content TEXT_INT '(or directly the textual content TEXT_INT if the lemmatisation step E20 has not been implemented) by using the semantic rules associated with the step PROBSTj of the model MOD (INSTj (C)) (step E220).

If the textual content TEXTJNT 'does not correspond to the semantic rules associated with the step PROBSTj of the MOD model (INSTj (C)) considered (answer no to the test step E230), the semantic analysis module 5C determines whether there exists a path instance associated with the client C and a model corresponding to this instance not yet examined (test step E240).

If there is still at least one client path instance and a corresponding unviewed customer path model (answer no to the test step E240), the semantic analysis module 5C chooses a new instance from the one not yet examined and the model corresponding customer journey (step E250 index incrementation j). Then it repeats the steps E220 to E240 using the semantic rules associated with the probable step of this new model.

If the TEXT_INT textual content corresponds to the semantic rules associated with the PROBSTj browse step of the MOD client journey model (INSTj (C)) (answer yes to the test step E230), the 5D module of the computing device 5 updates the instance INSTj (C) with the step of traversing PROBSTj (step E260). This update is reflected here by adding the PROBSTj browse step to the instance Instj (C) stored in the database 7. This step can also be associated with other contextual information as previously described. during the creation of the instance NEW_INST, as well as the communication channel used by the client C during the execution of the step PROBSTj.

Then it determines whether all the path instances associated with the client C and the corresponding customer journey patterns have been examined (test step E270). If there remain instances associated with the client C and models not yet examined (answer no to the test step E270), the semantic analysis module 5C chooses a new instance and the corresponding customer journey model among the instances associated with the client C not yet examined (step E250 index incrementation j), and repeats steps E220 to E270 on this new instance and this new model.

If, during the test step E240 or during the test step E270, the semantic analysis module 5C determines that all the instances associated with the client C and the corresponding client journey models have already been examined (yes answer to test step E240 or in the test step E270), phase PH2 is completed (step E280).

It is noted that this phase PH2 is all the more relevant that the INT interaction notified in step E10 is not the first INT1 interaction performed by client C as part of its CurrI customer journey, but a subsequent interaction ( eg INT2 or INT3 interaction).

Thanks to the implementation of PHI and PH2 phases, the IT device can reconstruct the customer C CurrI customer journey, regardless of the communication channel used during the journey steps made during this journey, even in the absence structured data to directly identify these stages of travel made via different communication channels. Note that if such structured data is available for an INT interaction, the path instance associated with the client C in the database 7 can be updated directly from these structured data.

As mentioned above, the contents of the databases 6 and 7 can advantageously be consulted thanks to the aforementioned graphic interfaces 5E.

Claims (14)

An automatic method for determining a customer path in a multichannel communication system, said method being adapted to be implemented by a computing device (5) and comprising: a reception step (E10), originating from a communication application (4), a notification (NOTIF (INT)) of an interaction (INT) via a communication channel (CH) of a client (C) with an entity of the communication system (3, 4), this notification comprising at least one information (INFO_ID) for identifying the client and a textual content (TEXT_INT) characteristic of this interaction; A consultation step (E30, E180) of a first database (6) and obtaining (E40, E190) of at least one customer journey model, said first database comprising models of customer journeys preestablished, each pre-established customer journey model representing a sequence of determined journey steps, at least one journey step being associated with a predetermined set of semantic rules; A semantic analysis step (E60, E220), for at least one obtained customer journey model, textual content (TEXT_INT) with the semantic rules associated with at least one traversing step (STl_MOD1, ..., STl_MODN; PROBSTl, ... PROBSTM) of the obtained customer journey model; If at the end of the semantic analysis step, it is determined that at least one step of a journey of a customer journey model obtained corresponds to the textual content, a step of creation (E140) or setting day (E260), in a second database (7), at least one client journey instance (NEW_INST (C), INST1 (C), .., INSTM (C)) associated with an identifier of the client, this instance of course corresponding to a customer journey in progress by the customer. 2. Determination method according to claim 1 further comprising a step of consulting (E140) the second database (7) and obtaining at least one path instance associated with a client identifier, the semantic analysis step (E220) of the textual content being implemented with the semantic rules associated with at least one step of traversing a customer journey model (M01, ..., M0DM) corresponding to said at least one instance obtained by consulting the second database. 3. A method of determining according to claim 2 wherein the semantic analysis step (E220) comprises, for each instance of course (INSTj (C)) obtained by consulting the second database, the determination (E200), in using said customer journey model corresponding to this route instance and taking into account the last step of the journey made of this route instance, at least one probable journey step (PR0BST1, ... PR0BSTM) following the last step of the course, the textual content being analyzed semantically (E220) with the semantic rules associated with the said at least one probable path step. 4. Determination method according to any one of claims 1 to 3 wherein said at least one instance of customer journey created or updated during the creation or updating step corresponds to the same object as the interaction. notified by the communication application. 5. Determination method according to any one of claims 1 to 4 wherein at least one step of a path of a customer journey model is associated with at least one semantic concept defined by an ontology. 6. Determination method according to any one of claims 1 to 5, further comprising a step of lemmatization (E20) textual content (TEXT_INT) characteristic of the interaction preceding the step of semantic analysis, the step of semantic analysis being implemented on the textual content (TEXT_INT ') obtained after lemmatisation. 7. Determination method according to any one of claims 1 to 6 wherein the notification of the interaction further comprises at least one additional information among: - time information associated with the interaction, - an application identifier or the communication entity, - an identifier of an actor who participated in the interaction. 8. Determination method according to any one of claims 1 to 7 wherein the customer identifier is obtained from the information contained in the notification to identify the customer and a correspondence table (TAB). between client identifiers on different channels of the multichannel communication system. 9. Determination method according to any one of claims 1 to 8 wherein the step of updating (E260) a customer journey instance includes the addition in the instance of customer journey associated with the identifier. the customer of a journey step corresponding to the textual content. 10. Determination method according to any one of claims 1 to 9 wherein a path step corresponding to the textual content is a first step (STl_MODi) of a customer journey model obtained and the creation step (E130) comprises the creation of a customer journey instance associated with the customer identifier having said first step of the customer journey model obtained as a first step. A computer program (PROG) comprising instructions for performing the steps of the determining method according to any one of claims 1 to 10 when said program is executed by a computer. 12. A computer-readable recording medium (9) on which is recorded a computer program comprising instructions for carrying out the steps of the determining method according to any one of claims 1 to 10. 13. A computer device (5) configured to automatically determine a customer journey in a multi-channel communication system, said computing device comprising: a receiving module (5A) able to receive, from a communication application, a notification of an interaction of a client with an entity of the communication system via a communication channel, this notification comprising at least one piece of information making it possible to identify the client and a textual content characteristic of the interaction; A consultation module (5B), configured to consult a first database (7) and obtain at least one customer journey model, said first database comprising pre-established customer journey models, each pre-established customer journey model representing a sequence of determined path steps, at least one path step being associated with a predetermined set of semantic rules; A semantic analysis module (5C) configured to analyze, for at least one obtained customer journey model, the textual content with the semantic rules associated with at least one step of the obtained customer journey model; and a module (5D) for creating or updating, activated if the semantic analysis module determines that at least one step of the journey of a customer journey model obtained corresponds to the textual content, this module being configured to create or update in a second database (6), at least one customer journey instance associated with a customer identifier and corresponding to a customer journey in progress by the customer. 14. Computer system (1) comprising: - a first database (7) comprising models of customer journeys, each customer journey model representing a sequence of determined journey steps, at least one journey step being associated with a predetermined set of semantic rules; - A second database (6) comprising instances of customer journey, each path instance being associated with a customer identifier and corresponding to a customer journey in progress by the customer; and - a computing device (5) according to claim 13 able to consult the first database and to create or update at least one instance of customer journey in the second database.