Method, system, and article of manufacture for assigning values of a customer system to records of a contact system

Abstract

This disclosure relates to a technique for assigning values ​​in a customer record system to records in a contact record system. The method of this disclosure includes: receiving a first set of records from the contact record system; receiving a second set of records from the customer record system; creating a third set, which is a subset of the first set; creating a fourth set, which is a subset of the first set; creating a fifth set, which is a subset of the second set; creating a sixth set, which is a subset of the second set; creating a seventh set based on the third and fifth sets; creating an eighth set based on the fourth and sixth sets; training a model based on the seventh set to obtain a trained model; and creating a ninth set based on the eighth set and the trained model.

Description

Technical Field

[0001] This disclosure generally relates to analyzing contact center data, and more specifically, to a technique for assigning values ​​from a customer record system to records in a contact record system. Background Technology

[0002] A contact center is a system used to receive or transmit large volumes of contact information, such as voice telephone calls (or "contacts"), web text chat, email, and video calls. A contact center may include outbound contact centers, which generate a large volume of outgoing contacts from the contact center. Such outbound contact centers are typically used for applications such as selling products, collecting outstanding credit balances, or surveying consumer sentiment. A contact center may also include inbound contact centers, which receive a large volume of incoming contacts from customers. Such inbound contact centers are also used for selling or cross-selling products, servicing customer support or technical support inquiries, retaining customers, escalating and downgrading services, or for other applications.

[0003] A typical contact center maintains multiple data logging systems, such as a contact logging system and a customer logging system. The contact logging system can be a log generated by an Automated Contact Distribution (ACD) system, while the customer logging system can be a log generated by a Customer Relationship Management (CRM) system, where agents access customer information as a result of customer interactions. The ACD system may contain one or more records for each contact, including information identifying the specific contact. The CRM system may not create any records, or it may create one or more records for each contact, including information identifying the contact's outcome (such as sales or customer satisfaction). To optimize contact center operations, it may be necessary to combine ACD records with CRM records so that contacts in the ACD system can be correlated with outcomes in the CRM system. This combination requires a certain level of accuracy and can be complex and difficult to implement.

[0004] One method for accurately matching ACD records with CRM records is to assign a unique identifier to each record circulating in the ACD and CRM systems. However, unique identifiers may not be available in the ACD and CRM systems because not all contact centers maintain such unique identifiers for various reasons. In such cases, contact centers may employ alternative strategies to match ACD records with CRM records. U.S. Patent No. 11,399,096 discloses a strategy for matching ACD records with CRM records based on information such as timestamps, agent identifiers (“Agent IDs”), customer identifiers (“Customer IDs”), and other data.

[0005] However, this strategy can also be inaccurate. For example, ACD and CRM system records may be entered at different times, which diminishes the effectiveness of timestamps. Furthermore, the agent ID in the ACD system may differ from the identifier in the CRM system. Finally, the clocks of the ACD and CRM systems may be out of sync. These and other issues can make accurate integration difficult.

[0006] Therefore, an improved system is needed to combine ACD records with CRM records. Summary of the Invention

[0007] Embodiments of this disclosure provide a method for assigning values ​​in a customer record system to records in a contact record system, comprising: receiving a first set of records in the contact record system; receiving a second set of records in the customer record system; creating a third set, which is a subset of the first set; creating a fourth set, which is a subset of the first set; creating a fifth set, which is a subset of the second set; creating a sixth set, which is a subset of the second set; creating a seventh set based on the third set and the fifth set; creating an eighth set based on the fourth set and the sixth set; training a model based on the seventh set to obtain a trained model; and creating a ninth set based on the eighth set and the trained model; wherein the fourth set can be represented by a first matrix; wherein the ninth set can be represented by a second matrix; wherein the dimension of the second matrix is ​​based on the dimension of the first matrix; and wherein the vectors of the second matrix contain values ​​based on the sixth set.

[0008] Optionally, in the above method, the third group and the fifth group share a unique identifier.

[0009] Optionally, in the above method, the fourth group and the sixth group do not share any unique identifiers.

[0010] Optionally, in the above method, the seventh group is based on a precise combination of the third group and the fifth group.

[0011] Optionally, in the above method, the eighth group is based on a cross combination of the fourth group and the sixth group.

[0012] Optionally, in the above method, creating the ninth group based on the eighth group and the trained model includes: using the trained model to calculate the weights corresponding to the eighth group; and creating the ninth group according to the weights and the CRM attributes of the eighth group to generate a merged result, wherein the sum of the merged results in the ninth group does not exceed the sum of the values ​​of the CRM attributes in the sixth group.

[0013] Embodiments of this disclosure also provide a system for assigning values ​​in a customer record system to records in a contact record system, comprising: at least one computer processor configured to assign values ​​in the customer record system to records in the contact record system, wherein the at least one computer processor is further configured to: receive a first set of records in the contact record system; receive a second set of records in the customer record system; create a third set, which is a subset of the first set; create a fourth set, which is a subset of the first set; create a fifth set, which is a subset of the second set; create a sixth set, which is a subset of the second set; create a seventh set based on the third set and the fifth set; create an eighth set based on the fourth set and the sixth set; train a model based on the seventh set to obtain a trained model; and create a ninth set based on the eighth set and the trained model; wherein the fourth set can be represented by a first matrix; wherein the ninth set can be represented by a second matrix; wherein the dimension of the second matrix is ​​based on the dimension of the first matrix; and wherein the vectors of the second matrix contain values ​​based on the sixth set.

[0014] Optionally, in the above system, the third group and the fifth group share a unique identifier.

[0015] Optionally, in the above system, the fourth group and the sixth group do not share any unique identifiers.

[0016] Optionally, in the above system, the seventh group is based on a precise combination of the third group and the fifth group.

[0017] Optionally, in the above system, the eighth group is based on a cross-combination of the fourth group and the sixth group.

[0018] Optionally, in the above system, the at least one computer processor is further configured to: use a trained model to calculate weights corresponding to the eighth group; and create a ninth group based on the weights and the CRM attributes of the eighth group to generate a merged result, wherein the sum of the merged results in the ninth group does not exceed the sum of the values ​​of the CRM attributes in the sixth group.

[0019] Embodiments of this disclosure further provide an article of art for allocating values ​​in a customer record system to records in a contact record system, comprising: a non-transitory processor-readable medium; and instructions stored on the medium, wherein the instructions are configured to be readable from the medium by at least one computer processor, the computer processor being configured to allocate values ​​in the customer record system to records in the contact record system, thereby causing the at least one computer processor to operate to: receive a first set of records in the contact record system; receive a second set of records in the customer record system; create a third set, which is a subset of the first set; and create a fourth set. The process involves creating a third group, a subset of the first group; creating a fifth group, a subset of the second group; creating a sixth group, a subset of the second group; creating a seventh group based on the third and fifth groups; creating an eighth group based on the fourth and sixth groups; training a model based on the seventh group to obtain a trained model; and creating a ninth group based on the eighth group and the trained model. The fourth group can be represented by a first matrix; the ninth group can be represented by a second matrix; the dimension of the second matrix is ​​based on the dimension of the first matrix; and the vectors of the second matrix contain values ​​based on the sixth group.

[0020] Optionally, in the above-described articles, the third group and the fifth group share a unique identifier.

[0021] Optionally, in the above-described articles, the fourth group and the sixth group do not share any unique identifiers.

[0022] Optionally, in the above-described articles, the seventh group is based on a precise combination of the third group and the fifth group.

[0023] Optionally, in the above-described articles, the eighth group is based on a cross-combination of the fourth group and the sixth group.

[0024] Optionally, in the above-described article of manufacture, the at least one computer processor is further configured to: use a trained model to calculate weights corresponding to the eighth group; and create a ninth group based on the weights and the CRM attributes of the eighth group to generate a merged result, wherein the sum of the merged results in the ninth group does not exceed the sum of the values ​​of the CRM attributes in the sixth group.

[0025] Embodiments of this disclosure include a method for assigning values ​​to records in a contact record system, comprising: receiving a first set of records from the contact record system; receiving a second set of records from a customer record system; assigning each record in the first set of records to each record in the second set of records to create a third set of records; determining a fourth set of records, the fourth set of records being a score for each record in the third set of records; creating a fifth set of records based on the third set of records and the fourth set of records, the fifth set of records being a value corresponding to the first set of records; and assigning the fifth set of records to the first set of records to create a result data set record.

[0026] Optionally, in the above method, the values ​​from the fourth group of records are all less than 1; and the sum of the values ​​from the fourth group of records does not exceed 1.

[0027] Optionally, in the above method, the sum of the values ​​from the fifth group of records is equal to the sum of the values ​​from the second group of records.

[0028] Optionally, in the above method, the sum of the values ​​from the fifth group of records is less than the sum of the values ​​from the second group of records.

[0029] Optionally, in the above method, the contact record system is an automatic contact allocation system, and the customer record system is a customer relationship management system.

[0030] Optionally, the method further includes: determining a scoring method based on attributes from the first set of records and the second set of records.

[0031] Alternatively, in the above method, a scoring method for calculating the value from the fourth set of records is designed using a set of known exact matching records.

[0032] Optionally, the method further includes filtering the first group of records and the second group of records based on known matching attributes.

[0033] Optionally, the method further includes: using the resulting data set records to create a matching model between agents and customers.

[0034] Optionally, the method further includes: using the resulting data set records to optimize the pairing model.

[0035] Optionally, the method further includes: using ID-based matching before receiving the first set of records from the contact record system.

[0036] Optionally, the method further includes: re-evaluating the scoring method based on a set of known exact matching records.

[0037] Embodiments of this disclosure include a system for assigning values ​​to records in a contact record system, comprising: at least one computer processor configured to assign values ​​to records in the contact record system, wherein the at least one computer processor is further configured to: receive a first set of records from the contact record system; receive a second set of records from a customer record system; assign each of the first set of records to each of the second set of records to create a third set of records; determine a fourth set of records, the fourth set of records being a fraction of each of the third set of records; create a fifth set of records based on the third set of records and the fourth set of records, the fifth set of records being values ​​corresponding to the first set of records; and assign the fifth set of records to the first set of records to create a result data set record.

[0038] Optionally, in the above system, the values ​​from the fourth group of records are all less than 1; and the sum of the values ​​from the fourth group of records does not exceed 1.

[0039] Optionally, in the above system, the sum of the values ​​from the fifth group of records is equal to the sum of the values ​​from the second group of records.

[0040] Optionally, in the above system, the sum of the values ​​from the fifth group of records is less than the sum of the values ​​from the second group of records.

[0041] Optionally, in the above system, the contact record system is an automatic contact allocation system, and the customer record system is a customer relationship management system.

[0042] Optionally, in the above system, the at least one computer processor is further configured to determine a scoring method based on attributes from the first set of records and the second set of records.

[0043] Optionally, in the above system, a scoring method for calculating the value from the fourth set of records is designed using a set of known exact matching records.

[0044] Optionally, in the above system, the at least one computer processor is further configured to filter the first group of records and the second group of records based on known matching attributes.

[0045] Optionally, in the above system, the at least one computer processor is further configured to use the resulting data set records to create a matching model between agents and customers.

[0046] Optionally, in the above system, the at least one computer processor is further configured to optimize the pairing model using the result data set records.

[0047] Optionally, in the above system, the at least one computer processor is further configured to use ID-based matching before receiving the first set of records from the contact record system.

[0048] Optionally, in the above system, the at least one computer processor is further configured to re-evaluate the scoring method based on a set of known accurate matching records.

[0049] Embodiments of this disclosure include an article of manufacture for assigning values ​​to records in a contact record system, comprising: a non-transitory processor-readable medium; and instructions stored on the medium, wherein the instructions are configured to be readable from the medium by at least one computer processor, the at least one computer processor being configured to assign values ​​to records in the contact record system, such that the at least one computer processor performs the following actions: receiving a first set of records from the contact record system; receiving a second set of records from a customer record system; assigning each of the first set of records to each of the second set of records to create a third set of records; determining a fourth set of records, the fourth set of records being fractions of each of the third set of records; creating a fifth set of records based on the third set of records and the fourth set of records, the fifth set of records being values ​​corresponding to the first set of records; and assigning the fifth set of records to the first set of records to create a result data set record.

[0050] Optionally, in the above-described article, the values ​​from the fourth set of records are all less than 1; and the sum of the values ​​from the fourth set of records does not exceed 1.

[0051] Optionally, in the above-described article, the sum of the values ​​from the fifth set of records is equal to the sum of the values ​​from the second set of records.

[0052] Optionally, in the above-described article, the sum of the values ​​from the fifth set of records is less than the sum of the values ​​from the second set of records.

[0053] Optionally, in the above-described product, the contact record system is an automatic contact allocation system, and the customer record system is a customer relationship management system.

[0054] Optionally, in the above-described article, the at least one computer processor is further configured to determine a scoring method based on attributes from the first set of records and the second set of records.

[0055] Optionally, in the above-described article, a scoring method for calculating the value from the fourth set of records is designed using a set of known, accurately matching records.

[0056] Optionally, in the above-described article, the at least one computer processor is further operated to filter the first group of records and the second group of records based on known matching attributes.

[0057] Optionally, in the above-described article of manufacture, the at least one computer processor is further operated to: use the resulting data set records to create a matching model between agents and customers.

[0058] Optionally, in the above-described article of manufacture, the at least one computer processor is further operated to optimize the pairing model using the resulting data set records.

[0059] Optionally, in the above-described article of manufacture, the at least one computer processor is further configured to use ID-based matching before receiving the first set of records from the contact record system.

[0060] Optionally, in the above-described article, the at least one computer processor is further operated to: re-evaluate the scoring method based on a set of known accurate matching records. Attached Figure Description

[0061] Figure 1 This is a schematic diagram illustrating a contact center system according to an embodiment of the present disclosure.

[0062] Figure 2 This is a flowchart illustrating a method for assigning values ​​in a customer record system to records in a contact record system according to an embodiment of the present disclosure.

[0063] Figure 3 This is a flowchart illustrating a method for assigning values ​​in a customer record system to records in a contact record system according to an embodiment of the present disclosure.

[0064] Figure 4 This is a schematic diagram illustrating a method for assigning values ​​in a customer record system to records in a contact record system according to another embodiment of the present disclosure.

[0065] Figure 5 This is a schematic diagram illustrating a comparison between time-based combination and probabilistic attribution according to an embodiment of the present disclosure.

[0066] Figure 6 This is a schematic diagram illustrating a comparison between time-based combination and probabilistic attribution according to another embodiment of this disclosure.

[0067] Figure 7 A schematic diagram illustrating a scenario of a possible combination method according to an embodiment of the present disclosure. Detailed Implementation

[0068] To more clearly illustrate the technical solutions of this disclosure, the accompanying drawings required for the embodiments of the present invention will be briefly described below. Obviously, the following drawings only relate to some embodiments of the present invention. Based on these drawings, those skilled in the art can obtain all other embodiments without creative effort.

[0069] As used herein, the term "module" can be understood to refer to computing software, firmware, hardware, and / or various combinations thereof, which can be configured as components of network elements, computers, and / or systems. A module should not be construed as software not implemented in hardware or firmware or not recorded on a processor-readable storage medium. These modules can be combined, integrated, separated, and / or replicated to support a variety of applications. These modules can be implemented on multiple devices and / or other components, either locally or remotely. Furthermore, these modules can be removed from one device and added to another, and / or contained in two devices.

[0070] The term “unique identifier” or “unique ID” as used in this article can be understood as referring to a unique call ID shared in the ACD and CRM systems.

[0071] The term "ACD attribute" as used in this article can be understood as referring to a characteristic or object that is closely related to or belongs to the ACD system. For example, an ACD attribute can be call duration, call start time, call end time, etc.

[0072] The term "CRM attribute" as used in this article can be understood as referring to a characteristic or object that is closely related to or belongs to the CRM system. For example, a CRM attribute could be order time, product type, etc.

[0073] Figure 1 This is a schematic diagram illustrating a contact center system according to an embodiment of the present disclosure. The contact center system 100 may include: a contact 110 arriving at the contact center system 100, an automatic contact distribution system 120, an agent 130 becoming available in the contact center system 100, and a customer relationship management system 140.

[0074] The ACD system 120 uses an algorithm to assign incoming calls to available agents. The ACD system 120 typically maintains a database that includes information such as the time the call was received at the contact center system 100, the agent interacting with the call, the duration of the call, and some identifiers of the customer initiating the call. The ACD system 120 may create one or more database records for each call, and each database record may have several fields, including: a call identifier, such as a billing phone number, customer ID, and email address; an agent identifier, such as agent ID and agent name; a timestamp of the date and the time of assignment; and other information related to the assignment.

[0075] CRM system 140 may contain a database of demographic information about individual customers, such as where the customer resides, when they became a customer, and the products and services they have currently purchased. CRM system 140 may also contain a database of historical customer interactions, such as records showing customer purchases or cancellations of previous purchases. As a result of interaction between agents and an agent desktop system (“agent desktop”), CRM system 140 may not create database records, may modify existing database records, or may create one or more new database records, wherein the agent desktop system is the system primarily used by agents to connect to other systems for information and configuration.

[0076] Many contact center systems 100 attempt to correctly correlate information in the ACD database with information in the CRM database. Associating CRM records (such as sales) with ACD records (such as calls) is called "linking." For example, contact center system 100 might want to determine whether a contact recorded in the ACD database resulted in a product purchase recorded in the CRM database. If the two systems share a unique identifier, precise linking between the two databases can be allowed. By doing so, contact center system 100 can seek to improve its performance in a variety of ways. For example, it can enhance its ability to evaluate its agents' performance or improve the routing strategy of its ACD system 120 in assigning contacts 110 to agents 130.

[0077] However, when there is a lack of shared unique identifiers between the two systems, it becomes difficult to accurately combine CRM records with ACD records. In such cases, several combination strategies can be considered. For example, U.S. Patent No. 11,399,096 discloses that the correlation between interaction events and result events can be determined by analyzing the interaction event time and result event time. This combination, also known as "time-based combination," combines ACD records to CRM records based on the time attributes of the records. For example, sales marked with specific timestamps are assigned to contacts whose start and end timestamps contain the timestamp of the sale. U.S. Patent No. 1,399,096 further describes a method that allows a data matching module to determine the probability that a given pair of only one CRM record and only one ACD record is a correct match, and to select the pair with the highest probability of correct match.

[0078] However, this time-based combination method can be inaccurate. While a specific CRM record should most likely be assigned to a specific ACD record, this is not definitive. Even though an ACD database record indicates that a specific agent received a contact from a specific customer within a specific time period, and a CRM database record indicates that a specific agent assigned a sale to a specific customer at a certain point in time within that period, the CRM record may still fail to be correctly assigned to the ACD record. Among other reasons, possible causes include a missynchronization of the system clocks between the CRM and ACD records, or a delay in an agent in agent 130 entering a sales entry for a contact in contact 110 into the CRM database.

[0079] If customer information in two databases is unrelated, the process of assigning ACD records to CRM records becomes less accurate. For example, ACD system 120 might contain a customer's phone number, while CRM system 140 might contain a different customer ID.

[0080] In light of the foregoing, a new and improved system may be needed to assign CRM information to ACD records. Among other optimizations, such a system would allow contact centers to evaluate, compensate, and train their agents to enhance ACD routing strategies for assigning contacts 110 to agents 130 and to improve the messaging used by agents 130 more accurately during their interactions with customers.

[0081] Figure 2 This is a flowchart illustrating a method for assigning values ​​in a customer record system to records in a contact record system according to an embodiment of the present disclosure, a method referred to as "probabilistic attribution" or "probabilistic combination".

[0082] The method disclosed herein includes: S201, receiving a first set of records from a contact record system; S202, receiving a second set of records from a customer record system; S203, creating a third set, which is a subset of the first set; S204, creating a fourth set, which is a subset of the first set; S205, creating a fifth set, which is a subset of the second set; S206, creating a sixth set, which is a subset of the second set; S207, creating a seventh set based on the third set and the fifth set; S208, creating an eighth set based on the fourth set and the sixth set; S209, training a model based on the seventh set to obtain a trained model; and S210, creating a ninth set based on the eighth set and the trained model. The fourth set can be represented by a first matrix; and the ninth set can be represented by a second matrix. The dimension of the second matrix is ​​based on the dimension of the first matrix; and the vectors of the second matrix contain values ​​based on the sixth set.

[0083] Probabilistic attribution is an improved assemblage method that takes into account the inherent uncertainty of the assemblage due to factors such as system clock asynchrony, and allows for a more accurate association between two systems.

[0084] Figure 3 This is a flowchart illustrating a method for assigning values ​​from a customer record system to records in a contact record system according to an embodiment of the present disclosure. The first group consists of ACD data received from records in the contact record system, and the second group consists of CRM data received from records in the customer record system. Based on whether precise combination between the ACD and CRM data is possible, the first group is divided into a third and a fourth subset, while the second group is divided into a fifth and a sixth subset. The third and fifth subsets can use precise combination, but the fourth and sixth subsets may not have available precise combination.

[0085] like Figure 3 As shown in the right half, the third and fifth groups are precisely combined, such as by unique ID, to obtain the seventh group, which is the known matching data. The seventh group is then used to train the outcome prediction engine (“OP engine”) using a prediction model. The trained OP engine receives candidate ACD records and candidate CRM records and returns the expected weighted data, calculated by the probability of using CRM records as candidate ACD records.

[0086] like Figure 3 As shown in the left half, the fourth and sixth groups (i.e., CRM and ACD records that cannot be precisely combined) are cross-combined. That is, each ACD record in the fourth group is combined with each CRM record in the sixth group, resulting in the eighth group, which represents all possible combinations between the fourth and sixth groups. The eighth group is then input into the OP engine as candidate ACD and CRM records to generate the ninth group, which is a matching data group that includes the value to be assigned and the corresponding record in the contact record system.

[0087] Figure 4 A schematic diagram illustrating a method for assigning values ​​in a customer record system to records in a contact record system according to another embodiment of the present disclosure is provided, and the method is explained. Figure 3 The data sets were obtained through the following steps.

[0088] Receive the first set of ACD records from the ACD system. Figure 4 (a) in the example, ACD1, ACD2, ... and ACDk, which are composed of call attributes. Figure 4 (a) in the table describes the ACD index, the availability of unique IDs for calls, and other call attributes. The first set of ACD records is then divided into two subsets, namely, a third set of ACD records with unique IDs. Figure 4 (c) in the middle, namely ACD1, ACD3, ... and ACDk) and the fourth group of ACD records that do not have a unique ID ( Figure 4 (d) in the first set, such as ACD2, ACD4, ... and ACDm). Regarding the availability of unique IDs in the first set, "T" indicates availability, while "F" indicates unavailability. Receive the second set of CRM records from the CRM system ( Figure 4 (b) in the example, CRM 1, CRM 2, ... and CRM j, which are composed of CRM attributes. Figure 4 Section (b) describes the CRM index, the availability of unique call IDs, and other CRM attributes. The second group is then divided into two additional subsets, namely, a fifth group of CRM records with unique IDs. Figure 4 (e) in the CRM series, such as CRM 2, CRM 5, ... and CRMj) and the sixth group of CRM records that do not have a unique ID. Figure 4 (f) in the second group, for example, CRM1, CRM3, ..., and CRMn). Regarding the availability of unique IDs in the second group, "T" represents available, while "F" represents unavailable. These divisions are based on whether each record has a unique identifier that allows for highly accurate combination across ACD and CRM systems.

[0089] The third and fifth groups have unique identifiers, therefore, ACD records in the third group can be precisely associated with CRM records in the fifth group. Based on the correlation between the ACD records in the third group and the CRM records in the fifth group, a seventh group can be created, such as... Figure 4 As shown in (g) in the diagram. Therefore, the seventh group is a precise combination of the third and fifth groups. The seventh group may contain ACD records consisting of call attributes from the third group and CRM records consisting of CRM attributes from the fifth group. The seventh group is then fed into the OP engine to "train" the OP engine (see [link to OP engine documentation]). Figure 4 (i)). The OP engine will use the seventh group, specifically the call attribute and CRM attribute, to create a predictive model that can receive candidate ACD records and candidate CRM records without unique identifiers and return the probability of a CRM record as a candidate ACD record.

[0090] The correlation between the ACD records in group four and the CRM records in group six cannot be easily determined because groups four and six do not share any unique identifiers. However, by applying a predictive model in the OP engine, the correlation between the ACD records in group four and the CRM records in group six can be predicted probabilistically.

[0091] like Figure 4 As shown in (h), the eighth group is created based on the fourth and sixth groups, for example, v 21v 23 v 24 ... v 2n ... v m1 v m3 v m4 ... and v mn Specifically, the eighth group is based on the cross-combination of the fourth and sixth groups; that is, each ACD record in the fourth group is combined with each CRM record in the sixth group. Each element of the eighth group can be a vector containing ACD call attributes and CRM attributes. For example, ACD 2 is assigned to CRM 1, CRM 3, ..., CRM n; ACD 4 is assigned to CRM 1, CRM 3, ..., CRM n; ...; ACD m is assigned to CRM 1, CRM 3, ..., CRM n, resulting in the eighth group matrix. Each item in the resulting matrix is ​​represented by a vector containing all the attributes of the corresponding ACD record and the corresponding CRM record. Each vector represents the probability that the corresponding CRM record is created as a direct result of the corresponding ACD record. In this example, v 21 Similarly, by combining the call attributes of ACD 2 with the customer relationship management attributes of CRM 1, v 41 Combine the call attributes of ACD4 with the customer relationship management attributes of CRM1. Furthermore, v 21 v 41 v 51 ... and v m1 They all contain customer relationship management attributes from the same CRM record (i.e., CRM 1).

[0092] Then, the OP engine ( Figure 4 In (i), a weighted reorganization can be created, corresponding to the elements of the eighth group. These weights can be labeled as w. xy This corresponds to the label of the element in the eighth group. For each element v in the eighth group... xy In other words, these weights can be based on the probability that ACD x leads to CRM y. For example, w 21 The probability that ACD2 leads to CRM1 can be based on this. To assign weights, the OP engine may require two inputs: the first is ACD-CRM “precise” combination data based on unique identifiers (Group 7, see...). Figure 4 (g) in the middle, and another is the ACD-CRM cross-linked data (Group 8, see...) Figure 4 (h)). The OP engine will use the ACD-CRM combination of the seventh group to train the prediction model, and then use the model to predict the weights of each ACD-CRM cross-combination record in the eighth group.

[0093] Subsequently, the OP engine generates a merged result for each ACD record in the eighth group based on the CRM entry and weight corresponding to that ACD record. For example, for ACD 2, the OP engine will accept input (v 21 v 23 ... v 2n ) and (w 21 w 23 ... w 2n The engine will accept input (v) and return the merged result value x2. Similarly, for ACD4, the OP engine will accept input (v) and return the merged result value x2. 41 v 43 ... v 4n ) and (w 41 w 43 ... w 4n The process calculates and returns the merged result value x4. This process may continue until a value is associated with each ACD record in the eighth group. The output of this process is the ninth group, where each ACD record in the fourth group has a corresponding merged result.

[0094] Figure 5 This is a schematic diagram illustrating a comparison between time-based attribution and probabilistic attribution according to an embodiment of this disclosure, where the agent ID is a unique identifier shared by the data ACD and CRM. In this case, agent 140 has 5 contacts: contact 1 starts at 9:00 AM and lasts for 9 minutes, contact 2 starts at 9:10 AM and lasts for 2 minutes, contact 3 starts at 9:13 AM and lasts for 5 minutes, contact 4 starts at 9:20 AM and lasts for 14 minutes, and contact 5 starts at 9:35 AM and lasts for 2 minutes, from 9:35 AM to 9:36 AM. At 9:21 AM, a sale of $50 is recorded in CRM system 140.

[0095] Time-based binding might group a sale with Contact 4 because the sale's timestamp falls within Contact 4's time window. However, an agent might discuss products and services with a customer at the start of a contact and then be marked as a sale later in the contact. If so, the sale is unlikely to occur at the start of Contact 4, as marked in the system. This could mean the CRM and ACD system clocks are out of sync. For example, if the CRM system clock is five minutes behind the ACD system clock, the sale would occur at 9:16 AM in the ACD system's clock. This would correctly place the sale at the end of Contact 3's time window, not at the beginning of Contact 4's time window.

[0096] The probability attribution method disclosed herein takes into account this likelihood and assigns the probability of a sale occurring to each contact that is reasonably close in terms of sales time. Probabilities of 56.25%, 12.50%, 25.00%, 6.24%, and 0.01% are assigned to contacts 1 through 5, respectively, based on attributes such as the sales timestamp and the duration of calls 1 through 5. Value attributions are calculated for contacts 1 through 5 based on the probability and value of the sale: contact 1 is attributable to $28.13 ($50 * 56.25%), contact 2 to $6.25 ($50 * 12.50%), contact 3 to $12.50 ($50 * 25.00%), contact 4 to $3.12 ($50 * 6.24%), and contact 5 to $0.00 ($50 * 0.01%). In this way, the sale has been allocated to each contact according to the probability that the contact has contributed to the sale.

[0097] Figure 6 This is a schematic diagram illustrating a comparison between time-based attribution and probabilistic attribution according to another embodiment of this disclosure, where the ACD system and the CRM system share a customer identifier. In this case, contact A is associated with two agents 141 and 142, where the contact paired with agent 141 begins at 9:00 AM and lasts for 9 minutes until 9:08 AM, while the contact paired with agent 142 begins at 9:18 AM and lasts for 2 minutes until 9:19 AM. At 9:18 AM, a sale of $40 is recorded in CRM system 140.

[0098] A time-based approach might combine the sales contact with agent 142, since the sales timestamp falls within the time window of the contact paired with agent 142. However, in reality, agent 141 might contribute more to the sales than agent 142 because agent 141 spends more time with the customer. After being initially persuaded by agent 141, the customer might simply call back to confirm the sale with agent 142. Therefore, simply combining the sales contact with agent 142 might be inaccurate, as agent 141 may also contribute to the sales, and their contribution might be greater than that of agent 142.

[0099] The probability attribution method of this disclosure takes into account this likelihood and assigns the probability of a sale occurring to each agent interaction that is reasonably close in terms of sales time. Based on characteristics such as the sales timestamp and the duration of the contact paired with agents 141 and 142, probabilities of 81.8% and 18.2% were assigned to the contacts paired with agents 141 and 142, respectively. Based on probability and sales value, the value attributable to the pairing with agent 141 was $32.72 ($40 * 81.8%), while the value attributable to the pairing with agent 142 was $7.28 ($40 * 18.2%). In this way, the sale has been allocated to each agent according to the probability that the agent contributed to the sale.

[0100] Figure 7 A schematic diagram illustrating a scenario of a possible combination method according to an embodiment of the present disclosure.

[0101] In some contact center systems 100, ACD data and CRM data are easily and accurately combined. This is likely because the two systems share unique data identifiers that link specific entries in the ACD database to specific entries in the CRM database, such as... Figure 7 Scenarios 1A to 1D are shown. Using a shared unique identifier, ACD records can be precisely linked to CRM records.

[0102] exist Figure 7 In Scenario 2, unique data identifiers are unavailable, but agent IDs and customer IDs exist in both systems. Therefore, based on agent IDs and / or customer IDs along with time parameters, ACD records can be accurately combined with CRM records. However, the preferred combination method remains probabilistic attribution.

[0103] exist Figure 7 In scenarios 3A and 3B, only one of the agent ID and customer ID exists in either system. Based on the time parameter and either the agent ID or the customer ID, ACD records can still be combined with CRM records. However, the preferred combination method is probabilistic attribution.

[0104] exist Figure 7 In scenario 4, the ACD system and the CRM system do not share any common agent or customer identifiers, and traditional combination methods may fail completely in this scenario. However, probabilistic attribution may succeed because the probabilistic attribution of this disclosure does not inherently require agent or customer identifiers for combination.

[0105] In summary, the probabilistic attribution method of this invention is applicable to almost all cases and is far more successful than other connection methods in scenarios where there is no precise binding.

[0106] The scope of this disclosure is not limited to the specific embodiments described herein. In fact, various other embodiments and modifications of this disclosure, besides those described herein, will be apparent to those skilled in the art from the foregoing description and drawings. Therefore, such other embodiments and modifications are intended to fall within the scope of this disclosure. Furthermore, although this disclosure has been described herein in the context of at least one specific implementation for at least one specific purpose in at least one specific environment, those skilled in the art will recognize that its usefulness is not limited thereto and that this disclosure can be advantageously implemented in any number of environments for any number of purposes.

Claims

1. A method for assigning values ​​in a customer record system to records in a contact record system, comprising: Receive the first set of records in the contact recording system, wherein the contact recording system includes ACD records; Receive a second set of records from the customer record system, wherein the customer record system includes CRM records; Create a third group, which is a subset of the records in the first group; Create a fourth group, which is a subset of the records in the first group; Create a fifth group, which is a subset of the records in the second group; Create a sixth group, which is a subset of the records in the second group; A seventh group is created based on the third group and the fifth group, wherein the ACD records in the third group and the CRM records in the fifth group can be precisely combined; An eighth group is created based on the fourth group and the sixth group, wherein the ACD records in the fourth group and the CRM records in the sixth group cannot be precisely combined, and each element of the eighth group is formed by cross-combining the ACD records in the fourth group and the CRM records in the sixth group. The model is trained based on the seventh group to obtain a trained model; and A ninth group is created based on the eighth group and the trained model; The fourth group can be represented by the first matrix; The ninth group can be represented by the second matrix; Wherein, the dimension of the second matrix is ​​based on the dimension of the first matrix; and The vector of the second matrix contains values ​​based on the sixth group; The step of creating the ninth group based on the eighth group and the trained model includes: The trained model is used to calculate the weights of each element in the eighth group; and A ninth group is created based on the weights and the CRM attributes of the eighth group to generate the merged result. The sum of the merged results in the ninth group shall not exceed the sum of the CRM attribute values ​​in the sixth group.

2. The method according to claim 1, wherein, The third group and the fifth group share a unique identifier.

3. The method according to claim 1, wherein, The fourth group and the sixth group do not share any unique identifiers.

4. The method according to claim 1, wherein, The seventh group is based on a precise combination of the third group and the fifth group.

5. A system for assigning values ​​in a customer record system to records in a contact record system, comprising: At least one computer processor is configured to assign values ​​from the customer record system to records in the contact record system, wherein the at least one computer processor is further configured to: Receive the first set of records in the contact recording system, wherein the contact recording system includes ACD records; Receive a second set of records from the customer record system, wherein the customer record system includes CRM records; Create a third group, which is a subset of the records in the first group; Create a fourth group, which is a subset of the records in the first group; Create a fifth group, which is a subset of the records in the second group; Create a sixth group, which is a subset of the records in the second group; A seventh group is created based on the third group and the fifth group, wherein the ACD records in the third group and the CRM records in the fifth group can be precisely combined; An eighth group is created based on the fourth group and the sixth group, wherein the ACD records in the fourth group and the CRM records in the sixth group cannot be precisely combined, and each element of the eighth group is formed by cross-combining the ACD records in the fourth group and the CRM records in the sixth group. The model is trained based on the seventh group to obtain a trained model; and A ninth group is created based on the eighth group and the trained model; The fourth group can be represented by the first matrix; The ninth group can be represented by the second matrix; Wherein, the dimension of the second matrix is ​​based on the dimension of the first matrix; and The vector of the second matrix contains values ​​based on the sixth group; The at least one computer processor is further configured to: The trained model is used to calculate the weights of each element in the eighth group; and A ninth group is created based on the weights and the CRM attributes of the eighth group to generate the merged result. The sum of the merged results in the ninth group shall not exceed the sum of the CRM attribute values ​​in the sixth group.

6. The system according to claim 5, wherein, The third group and the fifth group share a unique identifier.

7. The system according to claim 5, wherein, The fourth group and the sixth group do not share any unique identifiers.

8. The system according to claim 5, wherein, The seventh group is based on a precise combination of the third group and the fifth group.

9. An artifact for assigning values ​​in a customer record system to records in a contact record system, comprising: Non-transitory processor-readable medium; and The instructions stored on the medium The instructions are configured to be readable from the medium by at least one computer processor, the computer processor being configured to assign values ​​from the customer record system to records in the contact record system, thereby causing the at least one computer processor to operate such that: Receive the first set of records in the contact recording system, wherein the contact recording system includes ACD records; Receive a second set of records from the customer record system, wherein the customer record system includes CRM records; Create a third group, which is a subset of the records in the first group; Create a fourth group, which is a subset of the records in the first group; Create a fifth group, which is a subset of the records in the second group; Create a sixth group, which is a subset of the records in the second group; A seventh group is created based on the third group and the fifth group, wherein the ACD records in the third group and the CRM records in the fifth group can be precisely combined; An eighth group is created based on the fourth group and the sixth group, wherein the ACD records in the fourth group and the CRM records in the sixth group cannot be precisely combined, and each element of the eighth group is formed by cross-combining the ACD records in the fourth group and the CRM records in the sixth group. The model is trained based on the seventh group to obtain a trained model; and A ninth group is created based on the eighth group and the trained model; The fourth group can be represented by the first matrix; The ninth group can be represented by the second matrix; Wherein, the dimension of the second matrix is ​​based on the dimension of the first matrix; and The vector of the second matrix contains values ​​based on the sixth group. The at least one computer processor is further configured to: The trained model is used to calculate the weights of each element in the eighth group; and A ninth group is created based on the weights and the CRM attributes of the eighth group to generate the merged result. The sum of the merged results in the ninth group shall not exceed the sum of the CRM attribute values ​​in the sixth group.

10. The article of claim 9, wherein, The third group and the fifth group share a unique identifier.

11. The article of claim 9, wherein, The fourth group and the sixth group do not share any unique identifiers.

12. The article of claim 9, wherein, The seventh group is based on a precise combination of the third group and the fifth group.

Images