System for assigning an operator to an interaction

Abstract

System for assigning one of a plurality of operators to an interaction amongst one or more interactions each between a respective user and a user assistance centre comprising an artificial operator implemented on a computer, wherein each interaction comprises a real-time interaction between a user and the user assistance centre, the system comprising a controller configured to: obtain, during an information-collection phase of an interaction between a respective user and the user assistance centre, first information concerning a matter relating to the respective user; and determine an operator suitable to resolve the matter, based on said first information and on second information output by an AI-based operator assignment unit, the AI-based operator assignment unit being trained based on a dataset including at least the following information associated to each other: (i) past transcript information representing a transcript of an earlier interaction between a user and the user assistance centre; (ii) speech analysis information comprising parameters output from a speech analysis for the earlier interaction; and (iii) operator information identifying an operator assigned to said earlier interaction.

Description

[0001] 809 784 HOFFMANN EITLE S.R.L.

[0002] Milano

[0003] 1

[0004] Title of the invention:

[0005] System for assigning an operator to an interaction

[0006] TECHNICAL FIELD

[0007] The present invention relates to assigning operators to interactions between a user and a user assistance centre, and in particular to a system, method and a computer program.

[0008] BACKGROUND OF THE INVENTION

[0009] A user may contact an assistance centre via telephone, computer or any other telecommunication device and obtain remote assistance for resolving a specific matter. The contact may also be initiated by the assistance centre.

[0010] An assistance centre is typically manned by a plurality of human operators, which may form various groups to resolve a certain type of matters. For example, a group may be specialised in the technical assistance for a specific type of devices, or may be composed of operators skilled in a specific language, etc. In practice, the groups may be formed by assigning a specific tag to a profile associated with an operator (the associated profile may be stored in a database), the tag assigned to a profile indicating a skill of the operator.

[0011] A user interacts initially with an automated system that obtains information about the user's matter to be resolved, and assigns the user to a specific group of operators. Any operator of the group may then be individually assigned to the interaction. However, this approach fails to identify the operator that is most suitable to resolve a specific matter, which may lead to interactions remaining unresolved or taking longer to resolve a matter, thus causing an inefficiency in the use of network and computer resources of the assistance centre.

[0012] SUMMARY OF THE INVENTION

[0013] According to an aspect, the present invention aims to resolve or at least mitigate the problems related to the assignment of operators to interactions. 809 784 HOFFMANN EITLE S.R.L.

[0014] Milano

[0015] 2

[0016] Aspects of the invention can be described as follows:

[0017] Al. System for assigning one of a plurality of operators to an interaction amongst one or more interactions each between a respective user and a user assistance centre comprising an artificial operator implemented on a computer, wherein each interaction comprises a real-time interaction between a user and the user assistance centre, the system comprising a controller configured to:

[0018] - obtain, during an information-collection phase of an interaction between a respective user and the user assistance centre, first information concerning a matter relating to the respective user; and

[0019] - determine an operator suitable to resolve the matter, based on said first information and on second information output by an Al-based operator assignment unit, the Al-based operator assignment unit being trained based on a dataset including at least the following information associated to each other:

[0020] (i) past transcript information representing a transcript of an earlier interaction between a user and the user assistance centre;

[0021] (ii) speech analysis information comprising parameters output from a speech analysis for the earlier interaction; and

[0022] (iii) operator information identifying an operator assigned to said earlier interaction.

[0023] A2. System according to Al, wherein the dataset comprises, for each of the plurality of operators, information related to at least one earlier interaction assigned to the operator.

[0024] A3. System according to Al or A2, wherein the first information comprises transcript information representing a transcript of a conversation between the respective user and the artificial operator during the information-collection phase of the interaction.

[0025] A4. System according to any of Al to A3, wherein the first information comprises a question obtained from the user during the information-collection phase of the interaction. 809 784 HOFFMANN EITLE S.R.L.

[0026] Milano

[0027] 3

[0028] Optionally, in the system according to A4, at least one of the controller and the Al- based operator assignment unit is configured to determine one or more relevant parameters for the question out of a plurality of parameters, and to determine the suitable operator by prioritising the one or more relevant parameters.

[0029] A5. System according to any of Al to A4, wherein the Al-based operator assignment unit is configured to obtain the first information as an input and to output the second information based on the first information.

[0030] A6. System according to any of Al to A5, wherein the controller is configured to provide the first information to an interaction analytics system, wherein the interaction analytics system is configured to analyse the first information to identify third information related to the interaction, the third information indicating at least one of a sentiment, an emotion, a level of empathy, a keyword and a tendency of the interaction, wherein the controller is configured to determine the suitable operator based on the third information as well.

[0031] A7. System according to A6, wherein the Al-based operator assignment unit is configured to receive the third information as input and to output the second based on the third information.

[0032] A8. System according to A6 or A7, wherein the interaction analytics system comprises an Al-based interaction analytics unit.

[0033] A9. System according to any of A6 to A8, wherein the interaction analytics system comprises a speech analytics unit configured to analyse speech information comprised in the first information. 809 784 HOFFMANN EITLE S.R.L.

[0034] Milano

[0035] 4

[0036] A10. System according to any of Al to A9, wherein the Al-based operator assignment unit is configured to output, as second information or part thereof, operator information identifying an operator as suitable for the interaction.

[0037] All. System according to A10, wherein the Al-based operator assignment unit is configured to output, as second information or part thereof, an association between the operator information identifying the operator as suitable for the interaction and information indicating a level of correspondence between the operator's skills and the matter related to the user.

[0038] A12. System according to All, wherein the controller is configured to determine the operator suitable for the interaction taking into account the level of correspondence.

[0039] A13. System according to All or A12, wherein the controller is configured to determine the operator suitable for the interaction based on a plurality of factors comprising the second information, and wherein the controller is configured to, in the determination, assign a weight to the second information based on the level of correspondence.

[0040] A14. System according to any of Al to A13, wherein the dataset based on which the Al- based operator assignment unit is trained also includes, in association with the past transcript information, the speech analysis information and the operator information, outcome information indicating an outcome of the earlier interaction.

[0041] A15. System according to A14, wherein the outcome information indicates a classification of the earlier interaction, wherein the classification indicates at least whether the earlier interaction was successful or not.

[0042] A16. System according to any of Al to A15, wherein each of the plurality of operators is a human operator or an artificial operator implemented on a computer.

[0043] A17. System according to any of Al to A16, wherein the controller is configured to assign the determined operator to the interaction individually. 809 784 HOFFMANN EITLE S.R.L.

[0044] Milano

[0045] 5

[0046] A18. System according to any of Al to A17, wherein the dataset comprises a plurality of training information sets, wherein each training information set corresponds to a respective earlier interaction between a user and the user assistance centre, wherein the past transcript information, the speech analysis information and the operator information correspond to the respective earlier interaction.

[0047] A19. System according to A18, wherein the outcome information indicates an outcome of the respective earlier interaction.

[0048] A20. Method for assigning one of a plurality of operators to an interaction amongst one or more interactions each between a respective user and a user assistance centre comprising an artificial operator implemented on a computer, wherein each interaction comprises a real-time interaction between a user and the user assistance centre, the method comprising:

[0049] - obtaining, during an information-collection phase of an interaction between a respective user and the user assistance centre, first information concerning a matter relating to the respective user;

[0050] - determining an operator suitable to resolve the matter, based on said first information and on second information output by an Al-based operator assignment unit, the Al-based operator assignment unit being trained based on a dataset including at least the following information associated to each other:

[0051] (i) past transcript information representing a transcript of an earlier interaction between a user and the user assistance centre;

[0052] (ii) speech analysis information comprising parameters output from a speech analysis for the earlier interaction;

[0053] (iii) operator information identifying an operator assigned to said earlier interaction.

[0054] A21. Method according to A20, wherein the dataset comprises, for each of the plurality of operators, information related to at least one earlier interaction assigned to the operator. 809 784 HOFFMANN EITLE S.R.L.

[0055] Milano

[0056] 6

[0057] A22. Method according to A20 or A21, wherein the first information comprises transcript information representing a transcript of a conversation between the respective user and the artificial operator during the information-collection phase of the interaction.

[0058] A23. Method according to any of A20 to A22, wherein the first information comprises a question obtained from the user during the information-collection phase of the interaction.

[0059] Optionally, the method according to A18 comprises determining one or more relevant parameters for the question out of a plurality of parameters, and determining the suitable operator by prioritising the one or more relevant parameters.

[0060] A24. Method according to any of A20 to A23, wherein the second information is output by the Al-based operator assignment unit when the first information is used as an input to the Al-based operator assignment unit.

[0061] A25. Method according to any of A20 to A24, wherein the method comprises: analysing the first information to identify third information related to the interaction, the third information indicating at least one of a sentiment, an emotion, a level of empathy, a keyword and a tendency of the interaction, wherein the suitable operator is determined based on the third information as well.

[0062] A26. Method according to A25, wherein the second information is output by the Al-based operator assignment unit when the third information is used as an input to the Al-based operator assignment unit.

[0063] A27. Method according to A25 or A26, wherein an Al-based interaction analytics unit is used to analyse the first information to identify the third information.

[0064] A28. Method according to any of A25 to A27, wherein analysing the first information to identify the third information comprises analysing speech information comprised in the first information. 809 784 HOFFMANN EITLE S.R.L.

[0065] Milano

[0066] 7

[0067] A29. Method according to any of A20 to A28, wherein the second information comprises operator information identifying an operator as suitable for the interaction.

[0068] A30. Method according to A29, wherein the second information comprises an association between the operator information identifying the operator as suitable for the interaction and information indicating a level of correspondence between the operator's skills and the matter related to the user.

[0069] A31. Method according to A30, wherein the controller is configured to determine the operator suitable for the interaction taking into account the level of correspondence.

[0070] A32. Method according to A30 or A31, wherein the operator suitable for the interaction is determined based on a plurality of factors comprising the second information, and wherein a weight is assigned to the second information in the determination, based on the level of correspondence.

[0071] A33. Method according to any of A20 to A32, wherein the dataset based on which the Al- based operator assignment unit is trained also includes, in association with the past transcript information, the speech analysis information and the operator information, outcome information indicating an outcome of the earlier interaction.

[0072] A34. Method according to A33, wherein the outcome information indicates a classification of the earlier interaction, wherein the classification indicates at least whether the earlier interaction was successful or not.

[0073] A35. Method according to any of A20 to A34, wherein each of the plurality of operators is a human operator or an artificial operator implemented on a computer.

[0074] A36. Method according to any of A20 to A35, wherein the determined operator is assigned to the interaction individually. 809 784 HOFFMANN EITLE S.R.L.

[0075] Milano

[0076] 8

[0077] A37. Method according to any of A20 to A36, wherein the dataset comprises a plurality of training information sets, wherein each training information set corresponds to a respective earlier interaction between a user and the user assistance centre, wherein the past transcript information, the speech analysis information and the operator information correspond to the respective earlier interaction.

[0078] A38. Method according to A37, wherein the outcome information indicates an outcome of the respective earlier interaction.

[0079] A39. Computer program comprising instructions which, when executed by one or more processor, cause the one or more processor to carry out the method according to any of A20 to A38.

[0080] A40. Computer-readable storage medium storing the computer program according to A39.

[0081] BRIEF DESCRIPTION OF DRAWINGS

[0082] Embodiments of the present invention, which are presented for better understanding the inventive concepts, but which are not to be seen as limiting the invention, will now be described with reference to the figures in which:

[0083] Figure 1 shows a schematic diagram illustrating an example of a system for assigning an operator to an interaction between a user and a user assistance centre;

[0084] Figure 2 shows a schematic diagram illustrating an example of the system for assigning an operator to an interaction implemented in a user assistance centre connected to users;

[0085] Figure 3 shows a schematic diagram illustrating an example of an Al-based operator assignment unit being trained;

[0086] Figure 4 shows a schematic diagram illustrating an example of a trained Al-based operator assignment unit in operation;

[0087] Figure 5 shows processing operations implemented by a controller in an example implementation; 809 784 HOFFMANN EITLE S.R.L.

[0088] Milano

[0089] 9

[0090] Figure 6 shows processing operations for training an Al-based operator assignment unit in an example implementation;

[0091] Figure 7 shows a schematic diagram illustrating an example of a processing device that may be used to implement various elements of the system.

[0092] DETAILED DESCRIPTION

[0093] While exemplary embodiments will be described below, it will be apparent that various modifications can be made to these exemplary embodiments without departing from the broader spirit and scope of the invention. Therefore, the following description and the attached drawings should be considered illustrative and not restrictive.

[0094] Components described here such as the artificial operator, controller, Al-based operator assignment unit, interaction analytics system, Al-based interaction analytics unit, speech analytics unit, etc. can use any suitable communication link to exchange data, such as a wireless communication link (e.g. a Wi-Fi, a mobile phone data link such as LTE / 5G, Bluetooth or Bluetooth Low Energy (BLE)), a wired (e.g. DSL, fiber optic cable, Ethernet etc.). Each communication link may not be permanent.

[0095] Furthermore, each of these components can be created through hardware, software, or a combination thereof, as a single element or distributed in several elements. The same applies to the devices and / or systems described herein.

[0096] Numerous details are provided in the following description and in the attached figures in order to allow understanding of various implementation examples. However, it will be apparent to those skilled in the art that embodiments can be practiced without these details. For example, while the following figures and description may omit some connections between components, one skilled in the art will understand that such a connection may be present, even if not explicitly depicted or described. Likewise, the person skilled in the art directly derives that some of the described characteristics may be omitted. 809 784 HOFFMANN EITLE S.R.L.

[0097] Milano

[0098] 10

[0099] In existing user assistance centres, such as call centres, operators are grouped based on common skills and an operator from within the group is assigned to an interaction with a user. Whilst this approach simplifies the routing of a user to a specific operator and the management of operators (which can be managed by group), this fails to take into account the variation in skills between individual operators. A less suitable operator in a grouped may therefore be assigned to an interaction, risking that the user is not properly assisted with the matter, leading to a block in the resolution of the matter, or that the additional resources may need to be allocated to the resolution of the matter. In addition, such systems fail to take into account each operator's familiarity with the matter and its resolution.

[0100] As explained in the following, the present invention allows for a suitable operator to be determined taking into account the operator's individual suitability to resolve the matter, based on the operator's individual skills and experience with the matter (or similar matters).

[0101] With references to the figures, an example embodiment relating to a system for assigning one of a plurality of operators to an interaction amongst one or more interactions each between a respective user and a user assistance centre comprising an artificial operator implemented on a computer will be described. A customer who has purchased a certain product or service is an example of a user. A call centre is an example of a user assistance centre. The interaction can take place via text, voice, video call, or any combination of two or more of these. Other examples of the help centre are contemplated, such as an information centre, etc. The user assistance centre includes an artificial operator (or Al module, since in some examples it is implemented via Al) implemented via a computer. A computer is intended to mean any machine capable of processing information, both locally and in distributed manner, and through any combination of software and / or hardware. The artificial operator can be considered as an artificial agent created by computer. Optionally, the artificial operator is implemented via artificial intelligence, still preferably via generative artificial intelligence. Generative artificial intelligence means a technology capable of generating and / or interacting with a human through, for example, texts (e.g. messages) and / or voice conversations and / or images and / or videos. In particular, for interaction with 809 784 HOFFMANN EITLE S.R.L.

[0102] Milano

[0103] 11 humans, this may include the ability to manage natural languages thanks, for example, to the implementation of a large language model (LLM).

[0104] Figure 1 shows a schematic diagram illustrating an example of a system for assigning an operator to an interaction between a user and a user assistance centre.

[0105] The system 10 comprises a controller 100 and an Al-based operator assignment unit 110.

[0106] The system 10 may be implemented for a user assistance centre, such as a call centre. Users may contact the user assistance centre (or vice-versa) to resolve a matter relating to the user, and the system 10 may be used to assign a specific operator to each interaction, so the assigned operator interfaces with the corresponding user for resolving the matter related to that user.

[0107] Whenever a user starts interacting with the user assistance centre, the controller 100 is configured to obtain first information concerning a matter relating to the user, during an information-collection phase of the interaction, which is a phase where information is collected from the user, the collected information concerning a matter relating to the user. For example, the matter may be the reason for the interaction between the user and the user assistance centre. Although the information-collection phase is normally towards the beginning of the interaction, to arrive to a resolution of the matter quicker (i.e. it may be an initial phase of the interaction), this phase need not occur at the beginning of the interaction. For example, the interaction may include an earlier phase, such as a phase where the user is provided with information (e.g. general information or information which does require no or little input from the user).

[0108] The Al-based operator assignment unit 110 is used to determine an operator suitable to resolve the matter. Specifically, the Al-based operator assignment unit 110 is trained on a dataset which includes at least the following information associated with each other:

[0109] (i) past transcript information representing a transcript of an earlier interaction between a user and the user assistance centre; 809 784 HOFFMANN EITLE S.R.L.

[0110] Milano

[0111] 12

[0112] (ii) speech analysis information comprising parameters output from a speech analysis for the earlier interaction;

[0113] (iii) operator information identifying an operator assigned to said earlier interaction.

[0114] The Al-based operator assignment unit 110 may be trained using any suitable form of machine learning technique with a plurality of training information sets (i.e. the past transcript information, the speech analysis information and the operator information), each for a respective earlier interaction. Such interaction may be referred to as past interaction or earlier interaction in the sense that they occurred (or at least started, as they may still be ongoing) before the interaction for which an operator needs to be assigned. The Al-based operator assignment unit 110 may be trained using any known type of training, such as a supervised or unsupervised learning, to determine a feature indicative of a type of matter based on the past transcript information and / or speech analysis information, and determine whether the operator identified by the operator information should be considered suitable for the type of matter, for example by extracting a feature from the past transcript information and / or speech analysis information indicating whether the earlier interaction was successful.

[0115] The information sets may be collected from the interactions between the operators and users during a period of time before the system 10 is put into use, although information sets may continue to be collected whilst the system 10 is used (and the Al-based operator assignment unit may continue to be trained on newly collected information sets). By training the Al-based operator assignment unit 110 on the information from earlier interaction, the Al-based operator assignment unit can provide a prediction on how an operator will handle a new interaction.

[0116] The trained Al-based operation assignment unit 110 outputs second information for determining an operator suitable to resolve the matter, and the controller 100 is configured to determine, based on the first information (the information concerning the matter) and the second information, an operator suitable to resolve the matter. 809 784 HOFFMANN EITLE S.R.L.

[0117] Milano

[0118] 13

[0119] In some cases, two or more elements of the system 10 may be located on a same device. For example, the controller 100 and the Al-based operator assignment unit 110 may be located on the same device. In such cases, the system 10 may be defined as a device for assigning one of a plurality of operators to an interaction amongst one or more interactions each between a respective user and a user assistance centre comprising an artificial operator implemented on a computer.

[0120] Optionally, as a first example, the second information may include an indication of predicted success, which indicates for an operator how likely the operator (or each operator) will successfully resolve a matter for the user if the interaction is assigned to that operator. The second information may include a respective indication of predicted success for any number of operator, each predicting how the corresponding operator would (individually) resolve the matter. The controller 100 may therefore determine the suitable operator based on this indication of predicted success in the second information and the first information, for example by determining that an operator is suitable to resolve the matter if the indication of predicted success in the second information is a predetermined threshold or more. In cases where the second information includes the indication of predicted success for more than one operator, the controller 100 may determine the suitable operator as the one having the indication of predicted success with the highest value (i.e. indicating a most likely to successfully resolve the matter).

[0121] In some cases, the indication of predicted success may be associated with a specific type of matter, for example a matter having specific characteristics that may be determined from the past transcript information and / or the speech analysis information. In such cases, the second information may include, for any number of operator, different indications of predicted success for different types of matters. Thus, the controller 100 may first determine, based on the first information, what type of matter is to be resolved (for the interaction to be assigned to an operator), and then determine the suitable operator based on the indication of predicted success associated with the determined type of matter.

[0122] Optionally, as a second example, the second information may include an indication of a selected operator, i.e. information that identifies one of the plurality of operators selected 809 784 HOFFMANN EITLE S.R.L.

[0123] Milano

[0124] 14

[0125] (by the Al-based operator assignment unit) as the operator that is suitable for resolving the matter. This indication may correspond to the operator information on which the Al-based operator assignment unit was trained, or another identifier that the controller 100 can associate with an operator.

[0126] The indication of selected operator may indicate a single operator, or it may indicate more than one operators. In the latter case, the indication of selected operator may rank each operator in the subset (indicating that the operators are more or less suitable), or the operators in the subset may be indicated as equally suitable to resolve the matter.

[0127] Thus, the controller 100 may use the indication of selected operator and the first information to determine an operator as suitable to resolve the matter.

[0128] As with the indication of predicted success in the first example above, the indication of selected operator may be associated with a specific type of matter, and the second information may include, for any number of operator, different indications of selected operator(s) for different types of matters. Thus, the controller 100 may determine the type of matter to be resolved using the first information, and determine the indication of selected operator associated with that determined type of matter.

[0129] Optionally, the controller 100 may be configured to cause, once a suitable operator is determined, the suitable operator to be assigned to the interaction. For example, the controller 100 may generate one or more signals causing a terminal used by the determined operator to be connected to a terminal of the user and start interacting with the user.

[0130] In another example, the controller 100 may trigger a notification identifying the determined operator to another device in the user assistance centre 1 which instead determines which operator is to be assigned to a given interaction based on the notification.

[0131] Optionally, the controller 100 may be configured to cause the operator to be assigned to the interaction individually. In other words, once assigned to the interaction, the operator may be the sole entity with which the user interacts. Assigning the operator to the interaction 809 784 HOFFMANN EITLE S.R.L.

[0132] Milano

[0133] 15 individually allows the operator with greater control over the information that is provided to the user, and ensure that the operator has access to all of the information provided by the user during the interaction. It would be understood that assigning the operator to the interaction individually may still allow the operator to, if necessary later on, transfer the interaction to another, second operator. In case of such a transfer, the second operator may also be individually assigned to the interaction, and when transferred, the second operator may be provided access to all the information previously exchanged with the user during the interaction.

[0134] Optionally, the dataset may comprise, for each of the plurality of operators, information related to at least one earlier interaction assigned to the operator. For example, information may be collected for one or more interactions that were (or has been) handled by each operator. For each of these earlier interaction, transcript information and speech analysis information may be obtained, and associated with operator information identifying the operator assigned to that earlier interaction, to form part of the dataset on which the Al- based operator assignment unit is trained. Accordingly, the Al-based operator assignment unit is trained using data corresponding to each of the operators, such that the trained Al- based operator assignment unit can provide an estimate on whether each of the plurality of operators is likely to be able to resolve the matter for a new interaction.

[0135] Optionally, the first information comprises transcript information representing a transcript of a conversation between the respective user and the artificial operator during the information-collection phase of the interaction. For example, during the informationcollection phase, the sequence of information exchanged between the user and the artificial operator, such as questions and corresponding answers, may be collected in a text format. In cases where the artificial operator and the user interact using voice and / or video, a speech analysis may be performed to capture any information uttered by the user. Accordingly, the transcript information may provide information on the matter to be resolved. The transcript information may also provide additional context, such as a specific resolution sought by the user, a reference to a past interaction with the user assistance centre regarding the same matter, etc. This information in the transcript information helps the controller 100 better determine the operator that should be assigned to this interaction. 809 784 HOFFMANN EITLE S.R.L.

[0136] Milano

[0137] 16

[0138] Optionally, the first information comprises a question obtained from the user during the information-collection phase of the interaction. At least one of the controller and the Al- based operator assignment unit is configured to determine one or more relevant parameters for the question out of a plurality of parameters, and to determine the suitable operator by prioritising the one or more relevant parameters.

[0139] With reference to Figure 2, an example of an implementation of the system for assigning an operator in a user assistance centre will be described.

[0140] The system 10 as described above may be implemented as part of a user assistance centre 1 which also comprises a plurality 11 of operators (111 to 114), an artificial operator 12 and an interaction analytics system 13.

[0141] The user assistance centre 1 may be connected to a plurality of users 20, 21, 22 (and specifically, to a plurality of user device, i.e. devices used by respective users, such as computers, telephones, smartphones, etc.), via a network 3. Any suitable form of communication link may be established between the user assistance centre 1 and each user.

[0142] The artificial operator 12 is implemented on a computer, and may, for example, be implemented using artificial intelligence (Al). The artificial operator 12 may be configured to interface with each user 20, 21 and 22, to obtain data (such as text, audio and / or video data) relating to a respective matter concerning the corresponding user. The artificial operator 12 may also be configured to process data to interpret a message from the user (specifically, to interpret a content of the message) and to react to the user's message by generating a response to the message. This response is then transmitted to the user. The description omits for brevity details on the use of an Al in a user support center, as an interface to a user, which would become obvious to the skilled person based on the description. For example, the response generated by the artificial operator 12 can be a reaction to the user's message including, for example, answering a user's question, redirecting the conversation with the user, correcting the user, etc. The artificial operator and the user may interact with each other in resolving the issue by exchanging such messages. It is understood that the 809 784 HOFFMANN EITLE S.R.L.

[0143] Milano

[0144] 17 term "message" as used here refers to content and can be of any form, e.g., text, audio and / or video message. The data and / or audio / video signals may be part of a continuous transmission (e.g., a telephone or video call) or they may be separate transmissions that are interrupted. The user assistance centre 1 may comprise any number of artificial operator 12, each to interact with a respective number of users.

[0145] Each of the plurality 11 of operators 111 to 114 represent an operator that may interact with the user to resolve the matter concerning the user. Specifically, the user assistance centre may comprise, although not shown on Figure 2, communication equipment allowing data to be exchanged between operator devices provided respectively for each of the operators and the users 20, 21 and 22. The operators 111 to 114 may, through the corresponding operator device, interact with the user via text, voice and / or video.

[0146] Each operator may be a human (in which case the corresponding operator device may comprise input / output device allowing the human operator to visualise and / or hear information from the user) or an artificial operator, such as the artificial operator 12 described above. In the example provided on figure 2, operators 111, 112 and 113 are human operators, and operator 114 is an artificial operator. However, this is purely illustrative as there may be any number of two or more operators, and any combination of human and artificial operator is possible (including only human operators or only artificial operators). Each operator may, when resolving the matter, direct the user towards a certain resource of information, obtain information regarding the user, a service to which the user subscribed to, a device owned by the user, etc. (obtain information from the user or from a database connected to the user assistance centre) and provide information to the user to assist the resolution of the matter.

[0147] The interaction analytics system 13 is provided to obtain additional information regarding the interaction, to assist the determination of the suitable operator.

[0148] Specifically, the controller 100 may be configured to provide the first information to the interaction analytics system 13, and the interaction analytics system 13 may be configured to analyse the first information to identify third information related to the interaction. In 809 784 HOFFMANN EITLE S.R.L.

[0149] Milano

[0150] 18 other words, the third information is based on an analysis of the first information. The third information can indicate at least one of a sentiment, an emotion, a level of empathy, a keyword and a tendency of the interaction. The third information may thus provide an indication of the user's state during the interaction, and in particular during the informationcollection phase of the interaction. This may influence the determination of the operator suitable to resolve the matter. For example, if the third information indicate a negative sentiment or emotion of the user, such as frustration, despair, anger or confusion, an operator having a higher skill to establish empathic connection with the user, or an operator having a higher level of autonomy, or any other skill of the operator that may help reduce the user's frustration, may be more suitable. A similar differentiation may be made based on whether an interaction has a tendency towards a particular resolution of the matter or towards a deadlock (or any situation where the information provided by the artificial operator stops contributing to a resolution of the matter), etc, or if the tendency of the interaction shows a change in a sentiment, emotion, level of empathy.

[0151] The interaction analytics system 13 may then provide the third information to the system 10, or to a specific components thereof such as the controller 100 and / or the Al-based operator assignment unit 110.

[0152] Thus, the controller 100 may determine the suitable operator based on the third information as well. In other words, the determination of an operator suitable to resolve the matter may be based on the first information, on the second information output by an Al-based operator assignment unit, and on the third information provided by the interaction analytics system 13.

[0153] In some cases, the interaction analytics system 13 may comprise an Al module to identify the third information based on the first information. The Al module may an Al that is trained, via any suitable machine learning, or it may be a pre-configured rule engine or other suitable types of non-learning Al.

[0154] In the example shown on Figure 2, the interaction analytics system 13 comprises an Al-based interaction analytics unit 131 and a speech analytics unit 132. 809 784 HOFFMANN EITLE S.R.L.

[0155] Milano

[0156] 19

[0157] Optionally, the Al-based interaction analytics unit 131 may comprise an Al trained to process the first information (for example transcript information included in the first information) to detect whether the exchange of information between the artificial operator and the user during the information-collection phase indicate any of a sentiment, an emotion, a level of empathy, a keyword and a tendency of the interaction. The description omits for brevity details on the use of an Al to detect such information which would become obvious to the skilled person based on the description.

[0158] Optionally, the speech analytics unit 132 is configured to analyse speech information comprised in the first information. For example, the speech analytics unit 132 may identify that the first information indicates a certain tone, keyword, or other speech-related characteristic of the interaction with the user. For brevity, the description omits details on the use of an Al or on speech analysis which would become obvious to the skilled person based on the description.

[0159] Referring now to Figure 3, an example of training the Al-based operator assignment unit 110 will now be described.

[0160] The example of Figure 3 shows the Al-based operator assignment unit 110 as implementing a neural network, which may have an input layer, one or more intermediate layers and an output layer. However, it would be understood that the present invention is not limited to a specific type of Al, and may implement any suitable type of Al-based machine learning.

[0161] As shown on Figure 3, when training, the Al-based operator assignment unit 110 is coupled to a training module 36. The training module 36 causes the Al-based operator assignment unit 110 to receive an input and generate an output which indicates the suitability of an operator to resolve a particular matter. The training module then compares the output of the Al-based operator assignment unit 110 with a predetermined indication of suitability and trains the Al-based operator assignment unit 110 based on this comparison. For brevity, the present description omits details on the training of an Al, which would become obvious to the skilled person based on the description. 809 784 HOFFMANN EITLE S.R.L.

[0162] Milano

[0163] 20

[0164] Specifically, the dataset 30 comprises training information sets, each corresponding to a respective earlier interaction between a user and one of the plurality 11 of operators of the user assistance centre. Each set comprises a past transcript information 31 representing a transcript of an earlier interaction between a user and the user assistance centre, speech analysis information 32 comprising parameters output from a speech analysis for the earlier interaction, and operator information 33 identifying an operator assigned to said earlier interaction. The training module 36 uses each training information set in the dataset 30 to train the Al-based operator assignment unit 110.

[0165] Although not shown on Figure 3, the dataset 30 may comprise additional information associated with each earlier interaction (i.e. each training information set may comprise respective additional information). For example, the dataset 30 may comprise, for each earlier interaction, information derived from an analysis of the earlier interaction, for example by the interaction analytics system 13. This information derived from an analysis of the earlier interaction may correspond to the third information that the interaction analytics system 13 may obtain by analysing the past transcript information 31 and / or the speech analysis information 33. Accordingly, the Al-based operator assignment unit 110 may be trained to take into account additional information relating to the interaction, such as the third information described herein.

[0166] Optionally, the controller 100 may comprise the training module 36, for example the controller 100 may execute instructions of a computer program causing the controller 100 to perform the processing of the training module 36 as described herein.

[0167] In the example shown on Figure 3, each set of information also comprises, in association with the past transcript information, the speech analysis information and the operator information, outcome information 34 indicating an outcome of the earlier interaction. The outcome of the earlier interaction allows the earlier interaction to be classified into a category, and in particular may indicate, for example, whether the earlier interaction was successful (e.g. if a matter associated with the earlier interaction was successfully resolved during the earlier interaction) or not, whether at the conclusion of the earlier interaction, 809 784 HOFFMANN EITLE S.R.L.

[0168] Milano

[0169] 21 the user with which the user assistance centre interacted was satisfied (whether or not the matter was resolved), whether the earlier interaction was shorter than a threshold duration (e.g. a duration that may be predetermined for a type of matter with which the interaction was concerned) or whether it required the use of additional resources, etc.

[0170] Accordingly, with the outcome information, the earlier interaction may be classified into one of a positive training example or a negative training example, when the Al-based operator assignment unit 110 is being trained by the training module 36. Specifically, the training module 36 may, when inputting a set of information into the Al-based operator assignment unit 110, obtain the associated outcome information 34. The training module 36 may then obtain the second information 35 output by the Al-based operator assignment unit 110, and train the Al-based operator assignment unit 110 based on a comparison between the outcome information 34 with the second information 35.

[0171] A training information set labelled as a positive training example may cause the Al-based operator assignment unit 110 to reinforce an association between a type of matter indicated by the past transcript information 31 and / or the speech analysis information 32 in a training information set and the operator identified by the operator information 33 in the same training information set, thus leading to the operator being more likely to be identified by the Al-based operator assignment unit 110 as suitable when the Al-based operator assignment unit 110 receives, as an input, information indicating a same or similar matter for a subsequent interaction.

[0172] On the other hand, a negative training example may cause the Al-based operator assignment unit to weaken an association between the operator identified by the operator information 33 in a training information set and a type of matter indicated by the past transcript information 31 and / or the speech analysis information 32 in the training information set, thus leading to the operator being less likely to be identified as suitable to resolve a same or similar matter for a subsequent interaction.

[0173] Referring now to Figure 4, an example of a trained Al-based operator assignment unit will be described. 809 784 HOFFMANN EITLE S.R.L.

[0174] Milano

[0175] 22

[0176] As shown on Figure 4, the Al-based operator assignment unit 110 may receive, as an input, the first information 41 obtained by the controller 100.

[0177] In the example shown on Figure 4, the first information comprises two pieces of information from the information-collection phase of the interaction: the transcript information 411 representing the transcript of a conversation between the respective user and the artificial operator during the information-collection phase and the question 412 obtained from the user during the information-collection phase.

[0178] In addition, as shown on Figure 4, the Al-based operator assignment unit 110 may receive the third information 43 provided by the interaction analytics system 13.

[0179] The trained Al-based operator assignment unit 110 processes the input to generate, as an output, the second information 42, based on the first information and the third information.

[0180] In the example shown on Figure 4, the second information 42 includes operator information 421 identifying an operator as suitable for the interaction. For example, the operator information 421 may refer to the operator information 33 included in the dataset 30 for one of the earlier interaction on which the Al-based operator assignment unit 110 was trained. Accordingly, the controller 100 receiving the operator information 421 may determine which operator the Al-based operator assignment unit 110 identifies as suitable to be assigned to the interaction.

[0181] In the example shown on Figure 4, the second information also includes an association 422 between the operator information identifying the operator as suitable for the interaction and information indicating a level of correspondence between the operator's skills and the matter related to the user. Accordingly, the controller 100 may take into account the level of correspondence identified by the Al-based operator assignment unit 110 as a level of certainty for the information provided by the Al-based operator assignment unit 110 (the higher the level of correspondence the more likely the operator is deemed suitable to resolve the matter). In other words, for example, the controller 100 may be configured to 809 784 HOFFMANN EITLE S.R.L.

[0182] Milano

[0183] 23 determine the operator suitable for the interaction taking into account the level of correspondence. For example, the second information 42 may be one of many factors taken into account by the controller 100 to determine the operator suitable to resolve the matter. Each of these factors in the determination may be assigned a corresponding weight. For the case of the second information 42, the level of correspondence may be used to adjust the weight (e.g. the higher the level of correspondence, the higher the weight). In other words, for example, the controller 100 may be configured to determine the operator suitable for the interaction based on a plurality of factors comprising the second information, and the controller 100 may be configured to assign in the determination a weight to the second information based on the level of correspondence.

[0184] Referring now to Figure 5, processing operations implemented by a controller in an example implementation will now be described, when an interaction with a user is initiated and the user has been interfaced with the artificial operator during an information-collection phase.

[0185] At step 501, the controller 100 obtains first information 41 concerning a matter relating to the respective user, during an information-collection phase of an interaction between a respective user and the user assistance centre 1.

[0186] The controller 100 may obtain the first information concurrently with the informationcollection phase of the interaction, as information is being exchange with the user, or it may obtain the first information 41 at the end of the information-collection phase of the interaction (or when the information-collection phase nears its completion, to reduce any delay to the assignment of an operator to the interaction).

[0187] At step 502, the controller 100 provides the first information 41 to the interaction analytics system 13. The interaction analytics system 13 analyses the first information to identify third information 43 related to the interaction.

[0188] At step 503, the controller 100 obtains the third information 43 from the interaction analytics system 13. 809 784 HOFFMANN EITLE S.R.L.

[0189] Milano

[0190] 24

[0191] At step 504, the controller 100 provides the first information 41 and the third information to the Al-based operator assignment unit 110.

[0192] At step 505, the controller 100 obtains the second information 42 from the Al-based operator assignment unit 110.

[0193] At step 507, the controller 100 determines an operator suitable to resolve the matter based on the first information 41 and the second information 42.

[0194] At step 508, the controller causes the determined operator to start interacting with the user for resolving the matter.

[0195] As explained above, in some examples, the interaction analytics system 13 may be omitted. Accordingly, the processing of steps 502 and 503 shown on Figure 5 may be omitted as well. In addition, in such examples, the step 504 only involves providing the first information to the Al-based operator assignment unit 110.

[0196] In some examples, the controller 100 may not be responsible for causing the determined operator to be assigned to the interaction (and thus interact with the user). In other words, the processing of step 508 on Figure 5 may be omitted as well. Instead, the controller 100 may provide a notification indicating the operator that was determined to be suitable to resolve the matter.

[0197] Referring now to Figure 6, processing operations for training the Al-based operator assignment unit 110 in an example implementation will be described. Th

[0198] At step 601, a training information set corresponding to one earlier interaction from the dataset 30 is provided as an input to the Al-based operator assignment unit 110.

[0199] At step 602, an output of the Al-based operator assignment unit 110 is obtained, the output being generated based on the training information set that was input. 809 784 HOFFMANN EITLE S.R.L.

[0200] Milano

[0201] 25

[0202] At step 603, outcome information 34 from the training information set that was input into the Al-based operator assignment unit 110 is obtained.

[0203] At step 604, the Al-based operator assignment unit 110 is trained based on the output obtained at step 602 and the outcome information 34 obtained at step 603. For example, the training may be based on a comparison between the outcome information and the output.

[0204] More specifically, the training module 36 may determine whether the output of the Al-based operator assignment unit is in accordance with the outcome information (namely, whether the Al-based operator assignment unit output information indicating an operator for which the dataset holds a training information set with a positive outcome information) or not. If the output of the Al-based operator assignment unit is in accordance with the outcome information, the training module 36 may reinforce the association, and otherwise weaken the association between the operator information and the matter.

[0205] Referring now to Figure 7, a schematic diagram is described that illustrates an implementation of a device 70, such as the user device, or a device used by an operator, the artificial operator 12 and / or the controller 100. In a form of implementation, device 70 includes at least a processor 71, a memory 72 and an I / O interface 73. The processor 71 is configured to execute instructions included in a computer program stored on the memory 72 to perform any of the functions of device 70, as described here. The memory 72 is configured to store the computer program. In addition, memory 72 can also store additional information, which may include at the dataset for training the Al-based operator assignment unit 110, information for implementing the Al-based operator assignment unit 110, information for implementing the interaction analytics system 13 or part(s) thereof, any or all of the first information 41, second information 42 and third information 43, or any other information related to the interaction that may be exchanged between various entities, etc. During the execution of computer program instructions, processor 71 receives information from other devices via I / O interface 73 and outputs information to other devices via I / O interface 73. It is understood that the various elements of the device 70 need not be placed in a single housing, as the box is purely illustrative. For example, processor 71 may correspond to a server located on a rack ("shelf"), I / O 73 may correspond to a router and / or 809 784 HOFFMANN EITLE S.R.L.

[0206] Milano

[0207] 26 switch connected to the server, and memory 72 may correspond to a cluster storage system ("groups") comprising multiple storage units.

[0208] MODIFICATIONS AND VARIATIONS

[0209] Many modifications and variations can be made to the example embodiments described above.

[0210] In an example described above, the system 10 comprises the Al-based operator assignment unit 110. However, the present invention is not limited in that aspect. In alternative examples, the system 10 may be coupled to the Al-based operator assignment unit which may be comprised in the user assistance centre (but outside the system 10) or located at a remote location from the user assistance centre (for example an Al-based operator assignment unit may be coupled to a plurality of user assistance centres).

[0211] In an example described above, the first information comprises transcript information representing a transcript of a conversation between the respective user and the artificial operator during the information-collection phase of the interaction. However, the present invention is not limited in that aspect. In alternative examples, the first information may comprise information indicative of the matter without including transcript information at all. For example, the artificial operator may be configured to identify characteristics defining the matter when interacting with the user during the information-collection phase of the interaction, and provide first information including an indication of these characteristics to the controller 100.

[0212] In an example described above, the system 10 is connected to a speech analytics system 13. However, the present invention is not limited in that aspect. In alternative examples, the system 10 may comprise such speech analytics system 13. In other alternative examples, the speech analytics system 13 may be omitted entirely (i.e. the speech analytics system is optional). 809 784 HOFFMANN EITLE S.R.L.

[0213] Milano

[0214] 27

[0215] In examples described above, the speech analytics system 13 comprises an Al-based interaction analytics unit 131 and a speech analytics unit 132. However, the present invention is not limited in that aspect. In alternative examples, the speech analytics system 13 may comprise only one of the Al-based interaction analytics unit 131 and the speech analytics unit 132, or not comprise any sub-element. For example, the speech analytics system 13 may be a single entity, implemented in hardware and / or software, which analyse the first information to identify third information related to the interaction, where the third information indicates at least one of a sentiment, an emotion, a level of empathy, a keyword and a tendency of the interaction.

[0216] In an example described above, the dataset comprises outcome information 34 indicating an outcome of the earlier interaction. However, the present invention is not limited in that aspect. In alternative examples, the Al-based operator assignment unit 110 may be trained without the outcome information 34. For example, the Al-based operator assignment unit 110 may be trained by labelling each earlier interaction in the dataset 30 (and the corresponding training information set) as a positive training example or a negative training example without reference to the outcome information.

[0217] In an example described above, the first information 41 received by the trained Al-based operator assignment unit 110 comprises both the transcript information 411 and the question 412. However, However, the present invention is not limited in that aspect. In alternative examples, the first information 41 may omit the transcript information 411 and / or the question 412. In other words, the first information 41 may comprise only the transcript information 411 (and not the question 412), only the question 412 (but not the transcript information 411), or neither (it may instead comprise other information concerning the matter obtained from the information-collection phase of the interaction).

[0218] In an example described above, the trained Al-based operator assignment unit 110 obtains both the first information and the third information as inputs. However, the present invention is not limited in that aspect. In alternative examples, the Al-based operator assignment unit 110 may receive only the first information or the third information as an 809 784 HOFFMANN EITLE S.R.L.

[0219] Milano

[0220] 28 input, and generate the second information based on the only one of the first information and the third information received as an input.

[0221] In an example described above, the second information 42 includes the operator information 421 and the association 422. However, the present invention is not limited in that aspect, as the second information 42 may include only the operator information 421 without the association 422 (in other words, the Al-based operator assignment unit 110 may not provide an indication of a level of correspondence.

[0222] In examples described above, the training of the Al-based operator assignment unit 110 is described in connection with the outcome information 34 being used as the label of the training information set. However, as explained herein, the outcome information 34 may be omitted from the training information set. In such cases, step 603 may be omitted and replaced with processing for determining whether the earlier interaction corresponding to the training information set is a positive training example or a negative training example (for example, by determining whether the past transcript information 31 and / or the speech analysis information 32 in the training information set satisfies one or more predetermined conditions). In such cases, the processing at step 604 may comprise: if the training information set is a positive training example, and the output of the Al- based operator assignment unit 110 indicates the operator identified by the operator information 33 in the training information set as suitable for resolving the matter, training the Al-based operator assignment unit 110 to reinforce this association (between the operator information 33 and the past transcript information 31 and / or the speech analysis information 32 in the training information set); and if the training information set is a negative training example and the output of the Al- based operator assignment unit 110 indicates the operator identified by the operator information 33 in the training information set as suitable for resolving the matter, training the Al-based operator assignment unit 110 to weaken this association.

[0223] There are described above examples of training the Al-based operator assignment unit 110 using a supervised training method. However, the present invention is not limited in that aspect. The training of the Al-based operator assignment unit 110 may be unsupervised, a 809 784 HOFFMANN EITLE S.R.L.

[0224] Milano

[0225] 29 combination of supervised and unsupervised (e.g. semi-supervised), a reinforced training, etc.

[0226] Although reference has been made to a user assistance centre comprising an artificial operator implemented on a computer, it would be apparent to the skilled person that the present invention may be implemented in a user assistance centre without such artificial operator. For example, a user may be connected to a first human operator during an information-collection phase of the interaction, and the controller 100 may obtain the first information contemporaneously with the first human operator interacting with the user. Then, the controller 100 may determine a second human operator suitable to resolve the matter relating to the user and may trigger a commutation from an interaction between the user and the first human operator to an interaction between the user and the second human operator. The controller 100 may also trigger a notification that the second human operator is suitable to resolve the matter to an entity controlling which human operator is to interact with the user. It would be clear that the user may for a period of time interact with both the first human operator and the second human operator, which would smooth the commutation and allow the proper handover, thus improving the user's experience.

[0227] Any of the systems set forth herein can be implemented in a single device or in a plurality of devices, and in any combination of software and / or hardware. Furthermore, what has been said with reference to a system applies equally to a method that carries out the phases of such a system while it operates (and vice versa). Furthermore, what has been described with reference to a system and / or a method also applies to a computer program comprising instructions designed to execute, when said program is executed on a computer, all the phases according to a corresponding method. Naturally, the above description of embodiments and examples applying the principles recognized by the inventors is reported only as an example of such principles and must therefore not be understood as a limitation of the scope of the patent claimed here.

Claims

809 784 HOFFMANN EITLE S.R.L.Milano30 CLAIMS1. System (10) for assigning one of a plurality (11) of operators (111 - 114) to an interaction amongst one or more interactions each between a respective user and a user assistance centre (1) comprising an artificial operator (12) implemented on a computer, wherein each interaction comprises a real-time interaction between a user (20 - 22) and the user assistance centre, the system comprising a controller (100) configured to:- obtain, during an information-collection phase of an interaction between a respective user and the user assistance centre, first information (41) concerning a matter relating to the respective user; and- determine an operator suitable to resolve the matter, based on said first information and on second information (42) output by an Al-based operator assignment unit (110), the Al-based operator assignment unit being trained based on a dataset (30) including at least the following information associated to each other:(i) past transcript information (31) representing a transcript of an earlier interaction between a user and the user assistance centre;(ii) speech analysis information (32) comprising parameters output from a speech analysis for the earlier interaction; and(iii) operator information (33) identifying an operator assigned to said earlier interaction.

2. System according to claim 1, wherein the dataset comprises, for each of the plurality of operators, information related to at least one earlier interaction assigned to the operator.

3. System according to claim 1 or 2, wherein the first information comprises transcript information (411) representing a transcript of a conversation between the respective user and the artificial operator during the information-collection phase of the interaction.

4. System according to any of claims 1 to 3, wherein the first information comprises a question (412) obtained from the user during the information-collection phase of the interaction.809 784 HOFFMANN EITLE S.R.L.Milano315. System according to any of claims 1 to 4, wherein the Al-based operator assignment unit is configured to obtain the first information as an input and to output the second information based on the first information.

6. System according to any of claims 1 to 5, wherein the controller is configured to provide the first information to an interaction analytics system (13), wherein the interaction analytics system is configured to analyse the first information to identify third information (43) related to the interaction, the third information indicating at least one of a sentiment, an emotion, a level of empathy, a keyword and a tendency of the interaction, wherein the controller is configured to determine the suitable operator based on the third information as well.

7. System according to claim 6, wherein the Al-based operator assignment unit is configured to receive the third information as input and to output the second based on the third information.

8. System according to any of claims 1 to 7, wherein the Al-based operator assignment unit is configured to output, as second information or part thereof, operator information (421) identifying an operator as suitable for the interaction.

9. System according to claim 8, wherein the Al-based operator assignment unit is configured to output, as second information or part thereof, an association (422) between the operator information identifying the operator as suitable for the interaction and information indicating a level of correspondence between the operator's skills and the matter related to the user.

10. System according to claim 9, wherein the controller (100) is configured to determine the operator suitable for the interaction taking into account the level of correspondence.809 784 HOFFMANN EITLE S.R.L.Milano3211. System according to claim 9 or 10, wherein the controller (100) is configured to determine the operator suitable for the interaction based on a plurality of factors comprising the second information, and wherein the controller (100) is configured to, in the determination, assign a weight to the second information based on the level of correspondence.

12. System according to any of claims 1 to 11, wherein the dataset based on which the Al-based operator assignment unit is trained also includes, in association with the past transcript information, the speech analysis information and the operator information, outcome information (34) indicating an outcome of the earlier interaction.

13. System according to claim 12, wherein the outcome information (34) indicates a classification of the earlier interaction, wherein the classification indicates at least whether the earlier interaction was successful or not.

14. System according to any of claims 1 to 13, wherein each of the plurality of operators is a human operator or an artificial operator implemented on a computer.

15. System according to any of claims 1 to 14, wherein the controller is configured to assign the determined operator to the interaction individually.

16. System according to any of claims 1 to 15, wherein the dataset (30) comprises a plurality of training information sets, wherein each training information set corresponds to a respective earlier interaction between a user and the user assistance centre, wherein the past transcript information (31), the speech analysis information (32) and the operator information (33) correspond to the respective earlier interaction, and, wherein, optionally, the outcome information (34) indicates an outcome of the respective earlier interaction.

Images