JP7323727B1 - Information processing system, computer program, and information processing method - Google Patents

Abstract

An object of the present invention is to propose an appropriate sales plan to a salesperson. A business plan is created according to one aspect of the present disclosure (S140). The business plan includes one or more business activities, and designates the type and execution date and time of each of the one or more business activities. For each of the one or more sales activities included in the sales plan, an estimate is calculated as to the likelihood of the salesperson agreeing to perform the corresponding sales activity (S170). Based on this estimated value, it is determined whether or not to propose the sales plan to the salesperson (S180). If it is determined to propose, information for proposing the business plan to the salesperson is output (S230). [Selection drawing] Fig. 4

Description

The present disclosure relates to an information processing system and an information processing method.

Techniques for supporting sales activities have been conventionally researched. For example, a technology that accumulates sales results and the history of sales activities related to those results as performance data, and proposes a sales plan that incorporates sales activities that have the potential to achieve high results based on the accumulated data. is already known.

The applicant has already disclosed a technology related to a planning system and method capable of improving the results of patrol activities for patrol activities that patrol a plurality of locations (see Patent Document 1, for example).

JP 2022-045830

Plans proposed by the system may be changed or rejected by sales personnel for various reasons. The system predicts results and selects plans to propose to salespeople on the premise that plans will be executed as planned, but changes in plans by salespeople may adversely affect results. be.

That is, even if the system proposes a plan for sales activities that are expected to yield high results to the salesperson, if the acceptance rate for each sales activity included in the plan is low, there is a possibility that the plan will not lead to actual results.

Therefore, according to one aspect of the present disclosure, it is desirable to provide a technology capable of proposing an appropriate plan for sales activities to the salesperson, taking into account the possibility that the salesperson may reject the sales activity based on the plan.

According to one aspect of the present disclosure, an information processing system is provided. The information processing system is a system for proposing plans for sales activities. The information processing system includes a creation unit, a calculation unit, a determination unit, and a proposal unit.

The creating unit is configured to create a business plan including one or more business activities and designating the type and execution date and time of each of the one or more business activities. The calculation unit is configured to calculate, for each of the one or more sales activities included in the sales plan, an acceptance estimate, which is an estimate of the likelihood of the salesperson agreeing to perform the corresponding sales activity.

The determination unit is configured to determine whether to propose the sales plan to the salesperson based on the acceptance estimate. The proposing unit is configured to output information for proposing the business plan to the salesperson when the judging unit judges to propose.

According to this information processing system, there is no need to propose a sales plan with a low degree of tolerance to the salesperson. Therefore, it is possible to propose an appropriate sales plan to the salesperson, taking into consideration the possibility that the salesperson may reject the sales activity based on the plan.

According to one aspect of the present disclosure, the creation unit creates a first business plan as the business plan, and further creates a second business plan different from the first business plan. . The calculation unit can calculate, as the acceptance estimate value, an estimated value regarding the possibility that the salesperson will approve the execution of the corresponding sales activity for each of the one or more sales activities included in the first business plan.

The determination unit can determine whether or not to propose the first business plan to the salesperson based on the acceptance estimate for the first business plan. The proposing unit can output information for proposing the first business plan when the judging unit judges to propose. The proposal unit may be configured to output information for proposing the second business plan to the salesperson when the determination unit determines not to propose.

According to this information processing system, when there is a high possibility that sales activities based on the first business plan will be rejected by the salesperson, another business plan can be proposed as the second business plan. Therefore, it is possible to propose a plan that is easily accepted by the salesperson.

According to one aspect of the present disclosure, the calculation unit calculates an acceptance estimate for the first business plan as the first acceptance estimate, and one or more business activities included in the second business plan , an estimate of the likelihood of the salesperson agreeing to perform the corresponding sales activity can be calculated as a second acceptance estimate.

If the decision department decides not to propose, the proposal department decides whether or not to propose the second business plan to the salesperson based on the second acceptance estimate, and if it decides to propose , to output information for proposing a second sales plan to the salesperson.

According to one aspect of the present disclosure, the calculation unit may be further configured to calculate a change estimate value, which is an estimate value regarding the possibility of the salesperson changing the business plan, with respect to a plurality of change patterns. The determination unit may be configured to determine whether to propose the sales plan to the salesperson based on the acceptance estimate and the change estimate.

There are multiple types of changes in the business plan. Among them, there is also a type of change that places a heavy load on the salesperson. Therefore, if changeability is taken into consideration, plan proposals to sales personnel can be made more appropriately.

According to one aspect of the present disclosure, the determination unit calculates an estimated value related to the change load of the business plan by the salesperson based on the change estimated value, and based on the acceptance estimated value and the estimated value related to the change load of the business plan, It may be configured to determine whether to propose a plan to a salesperson.

According to this information processing system, since the change load is taken into account, it is possible to more appropriately propose a plan to the salesperson.

According to one aspect of the present disclosure, in an information processing system in which the creation unit creates a first business plan as the business plan and further creates a second business plan different from the first business plan, A calculator may be configured as follows.

That is, the calculation unit calculates the first acceptance estimate, which is an estimate of the likelihood that the salesperson will approve the execution of the corresponding sales activity for each of the one or more sales activities included in the first sales plan. Further, it can be configured to calculate a first change estimate value, which is an estimate value related to the possibility of the salesperson changing the first business plan, with respect to a plurality of change patterns.

The calculation unit calculates a second acceptance estimate, which is an estimate of the likelihood that the salesperson will approve the execution of the corresponding sales activity for each of the one or more sales activities included in the second sales plan, Furthermore, it can be configured to calculate a second change estimate, which is an estimate of the possibility of the salesperson changing the second sales plan, with respect to a plurality of change patterns.

The determination unit may be configured to determine whether to propose the first sales plan to the salesperson based on the first acceptance estimate and the first change estimate.
If the judgment department decides not to propose, the proposal department decides whether to propose the second sales plan to the salesperson based on the second acceptance estimate and the second change estimate, and decides to propose. In some cases, it can be configured to propose a second business plan to the salesperson.

According to one aspect of the present disclosure, the proposal unit calculates an estimated value regarding the change load of the second business plan by the salesperson based on the second change estimated value, the second acceptance estimate and the second It may be configured to determine whether to propose the second business plan to the salesperson based on the estimated value of the change load of the business plan.

According to one aspect of the present disclosure, an information processing system may comprise a model builder configured to build a predictive model of changeability based on a history of changes to previously proposed business plans. The prediction model is a prediction model that outputs estimated values regarding the possibility of a salesperson changing the sales plan for multiple change patterns, based on the input of at least one type of sales activity and execution date included in the sales plan. obtain. The calculator may be configured to calculate the change estimate based on the predictive model.

By constructing a prediction model based on the change history and calculating the change estimated value, the change possibility can be estimated with high accuracy. Therefore, it is possible to more appropriately determine whether or not to propose the business plan to the salesperson.

According to one aspect of the present disclosure, an information processing system is configured to build a predictive model of acceptability based on a history of acceptance and rejection of each business activity included in previously proposed business plans. A model builder may be provided. The predictive model may be a predictive model that outputs an estimate of the likelihood that the salesperson will accept at least one sales activity based on inputs of at least one sales activity type and execution date. The calculator may be configured to calculate the acceptance estimate based on the predictive model.

By constructing a prediction model based on the history of acceptance and refusal and calculating the estimated acceptance value, it is possible to estimate the possibility of acceptance with high accuracy. Therefore, it is possible to more appropriately determine whether or not to propose the business plan to the salesperson.

According to one aspect of the present disclosure, a computer program may be provided for causing a computer to implement the functions of the information processing system described above. According to one aspect of the present disclosure, there may be provided a computer program for causing a computer to implement the functions of the creating unit, the calculating unit, the determining unit, and the proposing unit in the information processing system described above.

According to one aspect of the present disclosure, an information processing method may be provided. The information processing method may include creating a business plan including one or more business activities, the business plan specifying the type and execution date and time of each of the one or more business activities.

The information processing method may include, for each of the one or more sales activities included in the sales plan, calculating an acceptance estimate, which is an estimate of the likelihood of the salesperson agreeing to perform the corresponding sales activity. Information processing methods may include determining whether to propose a sales plan to a salesperson based on the acceptance estimate. The information processing method may include outputting information for proposing a business plan to a salesperson when it is determined that the proposal should be made. The information processing method may be computer-implemented.

FIG. 4 is an explanatory diagram of information exchange between an information processing system and a salesperson; 1 is a block diagram showing the configuration of an information processing system; FIG. It is a figure showing a structure of plan data. It is a flowchart showing the plan proposal process which the processor of an information processing system performs. 4 is a flowchart showing data reading processing executed by a processor; 4 is a flowchart showing model building processing executed by a processor; It is a flowchart showing the plan evaluation process which a processor performs. It is a flow chart showing plan acceptance rate calculation processing which a processor performs. It is explanatory drawing regarding calculation of an acceptance probability. FIG. 10 is a flowchart showing editing cost calculation processing executed by a processor; FIG. FIG. 11 is an explanatory diagram regarding calculation of an edit probability;

Exemplary embodiments of the present disclosure are described below with reference to the drawings.
An information processing system 1 according to the present embodiment shown in FIG. 1 is a system for supporting sales activities by sales personnel by proposing an execution plan for sales activities as a sales plan to the sales personnel.

Examples of sales activities include activities for contributing to sales of products and/or services of the company to which the salesperson belongs, and specifically include negotiations regarding products and/or services. Examples of business plans include execution plans for sales activities for multiple customers, including existing and prospective customers.

The information processing system 1 acquires salesperson information describing the attributes of each of a plurality of salespersons, and acquires the activity history of each salesperson. The activity history includes information representing the editing history of business plans and the results of business activities.

The information processing system 1 is configured to select and propose an appropriate business plan to a salesperson based on salesperson information and activity history from among a plurality of sales plans created based on predicted results.

Specifically, the information processing system 1 considers the possibility that the salesperson may refuse to perform at least some of the sales activities in the sales plan, and selects a sales plan that is easy for the salesperson to accept from among a plurality of sales plans. It is configured to make a proposal to a salesperson.

According to this embodiment, the salesperson can provide the information processing system 1 with salesperson information and activity history and receive a sales plan from the information processing system 1 through the information terminal 30 owned by the salesperson.

The information processing system 1 includes a processor 11, a memory 12, a storage 13, and a communication interface 19, as shown in FIG. Processor 11 is configured to execute processing according to a computer program recorded in storage 13 .

The memory 12 includes RAM and is used as a work area when the processor 11 executes processing. The communication interface 19 is configured to communicate with the information terminal 30 via the wide area network. The storage 13 is composed of, for example, a hard disk drive (HHD) or solid state drive (SSD).

The storage 13 stores data necessary for searching for an appropriate business plan. For example, the storage 13 stores, for each salesperson, a salesperson ID that is an identifier of the corresponding salesperson, the age, sex, years of sales experience of the corresponding salesperson, and salesperson information describing the sales ratio for each type of negotiation. do.

Business negotiations include multiple types of business negotiations such as business negotiations through visits and business negotiations through web conferences. The sales ratio by business negotiation type represents the ratio of each type of business negotiations in the past business negotiations conducted by the salesperson. In addition, the storage 13 stores information on the address and business hours of the customer corresponding to the business negotiation destination.

The storage 13 further stores plan data describing the sales plan proposed to a plurality of sales representatives. The plan data shown in FIG. 3 has activity data of the corresponding salesperson in association with the salesperson ID for each salesperson.

The activity data for each salesperson comprises, for each previously proposed sales plan, a record for each sales activity relating to one or more sales activities included in the corresponding sales plan. The record of each sales activity includes an identification value relating to the type of corresponding sales activity, the execution date and time of the sales activity, the presence/absence of a proposal from the information processing system 1, the editing history, and information describing the results. The type of business activity referred to here can be defined based on the target, content, and location of the business activity.

The edit history takes one of the values "keep", "delete", "replace", and "insert". "Hold" indicates that no plan changes have been made by the salesperson to the sales activity described by the record. A sales activity whose edit history is “hold” and whose execution date and time has passed means that the salesperson has accepted and executed the sales activity proposed by the information processing system 1 .

"Deleted" indicates that the sales activity described by the record has been deleted from the sales plan. In other words, it means that the salesperson has refused to execute the sales activity proposed by the information processing system 1 for the sales activity whose edit history is "deleted".

“Replacement” indicates that the sales activity described by the record has been replaced from the sales activity proposed by the information processing system 1 in the same time zone. "Insert" means that the sales activity described by the record is not the sales activity proposed by the information processing system 1, but the sales activity newly added by the salesperson. However, "insertion" does not include "replacement".

The result described in each record of the activity data takes a value of either "yes" or "no". "Yes" means that the corresponding sales activity has achieved results, for example, that the purpose of the sales activity has been achieved by the sales activity above a predetermined standard. According to one example, "yes" means that a sales contract with a customer has been concluded through the corresponding sales activity (contract closed).

"None" means that no result was obtained for the corresponding sales activity, for example, it means that the purpose of the sales activity could not be achieved by the sales activity above a predetermined standard. According to one example, "None" means that the sales contract with the customer was not concluded by the corresponding sales activity (non-contract).

Next, details of the plan proposal processing executed by the processor 11 of the information processing system 1 when a business plan provision instruction is input from the information terminal 30 will be described with reference to FIG.

When the plan proposal process is started, the processor 11 identifies the target salesperson for whom the business plan is proposed (S110). Hereinafter, the target salesperson for whom the sales plan is proposed is referred to as the target salesperson. The processor 11 can acquire information on the salesperson ID of the target salesperson from the information terminal 30 to identify the target salesperson.

After that, the processor 11 executes a data reading process including a step of reading the activity data of the target salesperson based on the information of the salesperson ID of the target salesperson (S120). An example of the data reading process (S120) is shown in FIG.

According to the data reading process illustrated in FIG. 5, the processor 11 determines whether or not the target salesperson's degree of system usage exceeds a reference value (S121). The degree of system utilization means the degree of utilization of the business plan proposal service provided by the information processing system 1 .

For example, processor 11 may determine system utilization based on the amount of target salesperson activity data stored in the plan data. According to one example, system utilization may be a value proportional to the amount of activity data for the target salesperson.

When the processor 11 determines that the system utilization exceeds the reference value (Yes in S121), it reads the activity data of the target salesperson from the storage 13 (S122), and sets the read activity data as reference data ( S123). The reference data corresponds to the activity data referred to in the processes after S130. After that, the data reading process ends.

When the processor 11 determines that the system utilization is below the reference value (No in S121), the processor 11 reads the activity data of a plurality of related salespersons, including the individual activity data of the target salesperson (S126).

The plurality of related sales representatives here may be all sales representatives whose activity data is stored in the storage 13, or a group of sales representatives similar to the target sales representative, including the target sales representative, instead of all sales representatives. may be

The processor 11 further reads the salesperson information of a plurality of related salespersons from the storage 13 (S127). The salesperson information represents attributes of the corresponding salesperson, as described above.
After that, the processor 11 sets the read-out related activity data of the plurality of salespeople as reference data (S128). Hereinafter, the salesperson corresponding to the execution subject of the sales activity indicated by the reference data is referred to as the reference salesperson. After that, the processor 11 ends the data reading process.

After finishing the data reading process (S120), the processor 11 executes the model building process (S130). In the model building process shown in FIG. 6, the processor 11 uses the reference data as teacher data to create an acceptance prediction model, which is a prediction model relating to the possibility that the target salesperson will consent to the execution of each sales activity included in the sales plan. Construct (S131). Further, the processor 11 uses the reference data as teacher data to construct an edit prediction model, which is a prediction model regarding the possibility of the target salesperson editing the business plan (S135). After that, the processor 11 terminates the model building process.

According to this embodiment, the acceptance prediction model is constructed by machine learning using a neural network. The processor 11 constructs an acceptance prediction model by learning parameters of a neural network, for example, using teacher data created from reference data. The edit prediction model is also constructed by machine learning using neural networks.

Examples of acceptance predictive models and edit predictive models will now be described. According to an example, the acceptance prediction model can be a prediction model using a Recurrent Neural Network (RNN) and is constructed by defining a state vector _ht as follows.

A function sigmoid() means a sigmoid function. The input to the acceptance prediction model is a vector sequence x _seq =(x ₁ , x ₂ , _{. .} . , x _t , . ).

The sales activity at time t means the t-th sales activity when T sales activities are arranged in chronological order. T is determined by the designer considering the time dependence of the salesperson's acceptance action. For example, T can be a number as small as a few. Alternatively, T can be defined on the basis of the number of sales activities per day.

The presence or absence of consent is represented by binary values of 0 and 1. That is, y _appr ={0,1}. The probability p _appr is the salesperson's acceptance probability p _t=T for the sales activity at time t=T. The acceptance probability p _appr is calculated based on the sales activity at a total of T points in time, including the point t=T and (T−1) discrete points temporally earlier.

Parameters θ=(W _xh , W _hh , W _{hy ,} b _xh , b _hh , b _hy ) of the acceptance prediction model are learned using teacher data (x _seq , y _appr ) created from reference data. The value y _appr in the training data is determined from the editing history indicated by the corresponding record of sales activity. When the edit history is "retained", it means that the corresponding business activity has been "accepted", and when the edit history is other than "retained", the corresponding business activity is not "accepted" means

According to model learning using N training data from n = 1 to n = N based on a group of records related to sales activities that have been performed included in the reference data, the parameter θ is, for example, the following objective function: learned to minimize.

The parameter θ is optimized based on the cross-entropy error. Here, _{yappr, pappr} with index n means values based on n-th teacher data among N teacher data.

To generalize, the acceptance prediction model is based on the function F (x _seq, θ ₁ ) including the learned model parameter θ = θ ₁ and the vector sequence x _seq as elements, the formula p _appr = F (x _{seq ,} θ ₁ ).

The activity feature vector x _t , which is a component of the vector sequence x _seq , includes an activity explanation component, which is a feature vector component related to the sales activity at time t, and a subject explanation component, which is a feature vector component related to the salesperson performing the sales activity. include.

The activity feature vector _xt includes, as activity explanation components, an identification value relating to the type of sales activity at time t, and the execution date and time of the sales activity, specifically, the day of the week and time period during which the sales activity is performed. The activity feature vector _xt may further include, as an activity description component, one or more values representing one or more of the negotiation participants, the purpose of the negotiation, and the means of transportation to the negotiation location.

The activity feature vector x _t further includes, as a subject explanation component, one or more representing one or more of the salesperson ID, age, gender, years of sales experience, and sales ratio by business negotiation type of the salesperson who conducts the sales activity. The value of is included in the element.

When the acceptance prediction model is built based on reference data containing only the target salesperson's activity data, no subject explanatory component is needed in the activity feature vector _xt .

When the acceptance prediction model is constructed based on the reference data including the activity data of the salesperson other than the target salesperson, the acceptance prediction model is based on the activity features including the subject explanation component that explains the attributes of the salesperson who actually performed the sales activity. It is learned using the vector _xt as teacher data. After constructing the acceptance prediction model, the acceptance probability p _appr of the target salesperson is calculated based on the vector sequence x _seq using the activity feature vector _xt containing the subject explanation component explaining the attributes of the target salesperson.

According to an example, an edit prediction model is also built based on a recurrent neural network (RNN), defining a state vector _ht as follows.

Here, softmax( ) means a softmax function. The probability p _t in the edit prediction model is a vector containing multiple elements. The softmax function is a function that converts the probability _pt so that the sum of the element components of the probability _pt becomes one.

The input to the edit prediction model is the vector sequence x _seq =(x ₁ ,x ₂ ,...,x _t ,...,x _T ), similar to the acceptance prediction model. In the present embodiment, a total of five editing types are defined in connection with construction of an editing prediction model: "retention", "deletion", "replacement", "insertion", and "non-insertion". These editing types represent types of editing actions for sales activities.

The edit type "retain" means that no plan change is made by the salesperson to the corresponding sales activity. That is, "hold" means that there is no editorial action by the salesperson on the corresponding sales activity.

The edit type "deletion" means an edit action of deleting the corresponding business activity and not adding a new business activity in the same time slot as the corresponding business activity. The editing type "replacement" means an editing action of deleting the corresponding business activity and adding a new business activity in the same time zone as the business activity.

The edit type "insert" means an edit act of adding a new business activity next to the corresponding business activity. "Non-insert" means that no editing action is performed to add a new business activity next to the corresponding business activity.

The editing history described in the activity data indicates one of the values "hold", "delete", "replace", and "insert" as described above. Sales activities of records whose edit histories are "retain", "delete", "replace" and "insert" respectively correspond to sales activities whose edit types are "retain", "delete", "replace" and "insert" respectively.

The edit type “hold” is binary represented by 0, 1 by parameter y ₁₁ ={0, 1}. Similarly, the edit type "delete" is expressed in binary by parameter y ₁₂ ={0, 1}. The edit type “replacement” is expressed in binary by parameter y ₁₃ ={0, 1}.

The time period between the operations of any of the editing types "retain", "delete", and "replace" is set to the parameter _y21 ={0, 1} representing the editing type "insert" and the opposite value of the parameter _y21 . A binary representation is given by a parameter y ₂₂ =1−y ₂₁ representing the editing type “non-insertion” shown.

According to this embodiment, y _com1 =(y ₁₁ , y ₁₂ , y ₁₃ ) and y _com2 =(y ₂₁ , y ₂₂ ) are defined as a set of editing types.

The edit prediction model is a first edit prediction model that predicts edit probabilities p _com1 =(p ₁₁ ,p ₁₂ ,p ₁₃ ) corresponding to y _com1 =(y ₁₁ ,y ₁₂ ,y ₁₃ ) as probabilities p _com include. The edit prediction model further includes a second edit prediction model that predicts edit probabilities p _com2 =(p ₂₁ ,p ₂₂ ) corresponding to y _com2 =(y ₂₁ ,y ₂₂ ) as probabilities p _com .

The first edit prediction model is learned based on a group of supervised data (x _seq , y _com1 ) based on a group of records relating to completed sales activities included in the reference data. A second edit prediction model is learned based on a set of teacher data (x _seq , y _com2 ) based on a set of records relating to completed sales activities included in the reference data. The teacher data values y _com1 and y _com2 are determined from the editing history indicated by the corresponding sales activity record. Teacher data can be created from records having edit histories of, for example, "retain", "delete", and "replace".

The element p ₁₁ of the edit probability p _com1 represents the probability that "hold" will be performed as an editing action for the sales activity at time t=T, the element p ₁₂ will represent the probability that "delete" will be performed, and the element p ₁₃ will be , representing the probability that a “replacement” is made.

The element _p21 of the edit probability p _com2 represents the probability that a new sales activity will be inserted in the time period between the sales activity at time t=T and the next sales activity, and the element _p22 is the probability that it will not be inserted. (1−p ₂₁ ).

To generalize, the first edit prediction model is based on the learned model parameter θ = θ ₂₁ and the function G (x _seq , θ ₂₁ ) containing the vector sequence x _seq as elements, the formula p _com1 = G It is represented by (x _seq , θ ₂₁ ). The second edit prediction model is based on the function G (x _seq , θ ₂₂ ) containing the learned model parameter θ = θ ₂₂ and the vector sequence x _seq as elements, the formula p _com2 = G (x _seq , θ ₂₂ ).

After completing the model construction process (S130), the processor 11 creates a plurality of business plans based on the predicted results (S140). Each business plan includes one or more business activities, and designates the type and date of execution of each business activity.

For performance prediction, the processor 11 can use the reference data as teacher data to build a performance prediction model. By inputting possible business plans into this prediction model, the results of each of the possible business plans are predicted, and a predetermined number of business plans are selected in descending order of predicted results as the plurality of business plans. can be created. A high result may correspond to, for example, a large number of successful sales activities or a large number of contracts.

After that, the processor 11 sets k=1 (S150), and selects the business plan with the k-th predicted result from among the plurality of business plans created in S140 as the business plan to be evaluated ( S160). Here, the first-ranked business plan is the business plan with the highest predicted result among the plurality of business plans.

In subsequent S170, the processor 11 executes the plan evaluation process shown in FIG. When the plan evaluation process is started, the processor 11 executes the plan acceptance rate calculation process shown in FIG. 8 in S171.

In the plan acceptance rate calculation process, the processor 11 selects one of the sales activities included in the sales plan to be evaluated (S310), and generates a vector sequence x _seq including the activity feature vector x _t=T related to the selected sales activity. , as input to the acceptance prediction model (S320).

The vector sequence x _seq includes activity feature vectors x _t for T−1 sales activities performed temporally before the selected sales activity. Among them, an activity feature vector _xt related to the sales activity of the previous plan, which is a business plan proposed temporally before the business plan to be evaluated, can be included.

The previous plan may be a plan that has already been executed by a salesperson or a plan before execution. For this reason, as for the sales plan already proposed to the target salesperson, the sales plan before execution can also be recorded in the plan data.

In subsequent S330, the processor 11 inputs the vector sequence x _seq generated in S320 to the acceptance prediction model generated in S131 to calculate the acceptance probability p _appr of the selected sales activity.

In subsequent S340, the processor 11 determines whether or not the acceptance probability p _appr has been calculated for all sales activities included in the sales plan within the evaluation target. If a negative determination is made here (No in S340), the processor 11 selects a business activity for which the acceptance probability p _appr has not been calculated in S310, and calculates the acceptance probability p _appr for the selected business activity (S310- S330).

In this way, the processor 11 generates a vector sequence x _seq for each of the sales activities included in the sales plan as shown in FIG. Calculate _appr . According to FIG _. 9, for sales activities A1, _A2 , .

If the processor 11 makes an affirmative determination in S340, in subsequent S350, the processor 11 calculates the plan acceptance rate P _appr based on the sum Σp _appr of the acceptance probabilities p _appr for each business activity included in the business plan.

The plan acceptance rate P _appr corresponds to a quantified degree of acceptance for the business plan. The planned acceptance rate P _appr may be the sum Σp _appr of the acceptance probabilities p _appr for each business activity (P _appr =Σp _appr ). For example, if the number of business activities in each of the plurality of business plans created in S140 is uniformly the same, the plan acceptance rate P _appr can be calculated as the total sum Σp _appr . Alternatively, the plan acceptance rate P _appr may be a value obtained by dividing the total sum Σp _appr by the number of sales activities included in the business plan, that is, the average of the acceptance probabilities p _appr for each business activity.

After calculating the plan approval rate _Pappr in S350, the processor 11 ends the plan approval rate calculation process. After completing the plan acceptance rate calculation process in S171 (see FIG. 7), the processor 11 calculates the degree of plan editing based on the reference data in subsequent S173.

Specifically, the processor 11 determines the number M1 of records having an edit history of one of "hold", "delete", "replace", and "insert" in the reference data. The processor 11 further determines the number M2 of records having an edit history of any one of "delete", "replace", and "insert" other than "hold" in the reference data. The processor 11 calculates M2/M1 as the degree of planned editing. That is, the degree of plan editing corresponds to the percentage of sales activities that are edited out of the group of sales activities shown in the reference data.

In subsequent S175, the processor 11 determines whether or not the planned editing degree is greater than the reference value. When the processor 11 determines that the planned editing degree exceeds the reference value (Yes in S175), the processor 11 executes the editing cost calculation process (S177) shown in FIG. 10, and then executes the process of S179. On the other hand, when determining that the planned editing degree is equal to or less than the reference value (No in S175), the processor 11 executes the processing of S179 without executing the editing cost calculation processing (S177).

In the editing cost calculation process shown in FIG. 10, the processor 11 selects one of the sales activities included in the sales plan to be evaluated (S410), and selects a vector sequence including an activity feature vector x _t=T related to the selected sales activity. x _seq is generated as input to the edit prediction model (S420), similar to the process at S320.

In subsequent S430, the processor 11 inputs the vector sequence x _seq generated in S420 to the first editing prediction model and the second editing prediction model generated in S135 to calculate the editing probability p _com1 and p _com2 are calculated.

In subsequent S440, the processor 11 determines whether or not the edit probabilities p _com1 and p _com2 have been calculated for all sales activities included in the sales plan within the evaluation target. If a negative determination is made here (No in S440), the processor 11 selects a sales activity for which the edit probabilities p _com1 and p _com2 have not been calculated in S410, and calculates the edit probabilities p _com1 and p _com2 for the selected sales activity. is calculated (S410 to S430).

In this way, the processor 11 generates a vector sequence x _seq for each of the sales activities included in the sales plan as shown in _FIG . _com2 is calculated.

In subsequent S450, the processor 11 sets the first edit sequence E _s1 =(E _s1 [1], E _s1 [2], . . . , E _s1 [m], . _s1 [M]) and the second edit sequence E _s2 for "insertion" and "non-insertion" = (E _s2 [1], E _s2 [2], ..., E _s2 [m], ..., E _s2 [M ]).

Here, M is the number of sales activities included in the sales plan to be evaluated. E _s1 [m] represents an edit type predicted to be performed for the m-th sales activity included in the business plan among the edit types "retain", "delete", and "replace". E _s2 [m] is predicted to be performed in the time period between the m-th sales activity included in the sales plan and the next sales activity among the editing types "insert" and "non-insert" Represents the edit type.

When the edit probability p _com1 =(p ₁₁ , p ₁₂ , p ₁₃ ) is calculated for the m-th sales activity, the "hold" probability is the probability that the m-th sales activity is retained without being edited. The editing type that indicates the highest probability among p ₁₁ , the “deletion” probability p ₁₂ that is the probability that the m-th sales activity will be deleted, and the “replacement” probability p ₁₃ that is the probability that the m-th sales activity will be replaced. is set to E _s1 [m].

When the editing probability p _com2 =(p ₂₁ , p ₂₂ ) is calculated for the m-th sales activity, a new sales activity is inserted between the m-th sales activity and the next sales activity by editing. Of the “insertion” probability p ₂₁ , which is the probability of being inserted, and the “non-insertion” probability p ₂₂ , which is not, the value of the editing type indicating the higher probability is set to E _s2 [m].

In subsequent S460, the processor 11 calculates the plan editing cost _CE , which is the editing cost for the business plan to be evaluated, based on the first editing series _Es1 and the second editing series _Es2 . After that, the edit cost calculation process ends.

The plan editing cost _CE is a numerical value of the editing load of the sales plan to be evaluated by the target salesperson. In S460, the editing cost C[m] for the m-th sales activity is calculated by a predetermined function H(E _s1 [m], E _s2 [m]), and the editing cost C [m ] can be calculated as the planned editing cost _CE .

According to one example, in calculating the planned editing cost _CE , a score corresponding to the editing load can be defined for each of the editing types of "hold", "delete", "replace", "insert", and "non-insert". For example, depending on the editing load, 0 points for "keep", 1 point for "delete", 2 points for "replace", 0 points for "not insert" Five points can be defined for the point "Insert".

The editing cost C[m] may be the sum of the editing type score corresponding to E _s1 [m] and the editing type score corresponding to E _s2 [m]. The planned editing cost _CE can be the sum of the editing costs C[m] for m=1 to M thus calculated.

After the editing cost calculation process (S177) ends, the processor 11 calculates the plan tolerance Z based on the plan approval rate _Pappr and the plan editing cost _CE calculated in S171 (S179). After that, the processor 11 terminates the plan evaluation process.

The plan tolerance Z represents the tolerance of the sales plan by the target salesperson. The planning tolerance Z can be calculated according to a previously prepared function R( _Pappr , C _E ) (Z=R(P _appr , C _E )). The plan tolerance Z can be calculated as a higher value as the plan acceptance rate P _appr is higher and as the plan compilation cost _CE is lower.

For example, the plan tolerance Z can be calculated as a value proportional to the plan acceptance rate _Pappr and inversely proportional to the plan compilation cost _CE . For example, the planning tolerance Z can be calculated as a value proportional to ( _Pappr /C _E ).

In the process of S179, which is executed when a negative determination is made in S175, the processor 11 can set the plan editing cost _CE to a fixed value and calculate the plan tolerance Z. FIG. For example, processor 11 may set plan edit cost _CE to a value of 1 and calculate plan tolerance Z as a value proportional to P _appr .

After completing the plan evaluation process (S170), the processor 11 determines whether or not the plan tolerance Z exceeds a predetermined value (S180). If it is determined that the plan tolerance Z exceeds a predetermined value (Yes in S180), the processor 11 determines the k-th business plan as the business plan to be proposed to the target salesperson (S190). Plan proposal data including information explaining the is output to the target salesperson (S230).

At S230, the processor 11 can transmit the plan proposal data to the information terminal 30 of the target salesperson. At S230, the processor 11 can further register a record of each sales activity based on the proposed sales plan in the plan data. Since the sales activity corresponding to the record to be registered has not yet been executed at the time of registration, the result information is empty in the corresponding record. Similarly, the edit history information may be empty in the corresponding record. Alternatively, the edit history may have provisional information of "hold".

The information terminal 30 can display the sales plan proposed by the information processing system 1 to the target salesperson based on the plan proposal data received from the information processing system 1 . The target salesperson can edit the sales plan through the information terminal 30 . This edit history is transmitted to the information processing system 1 as part of the activity history described above and managed within the plan data. The editing history in the plan data is updated according to the editing operation of the target salesperson.

When the processor 11 determines in S180 that the plan tolerance Z is equal to or less than the predetermined value (No in S180), among the plurality of business plans created in S140, there is a business plan that has not been selected as an evaluation target in S160. It is determined whether or not there is (S200).

Here, if it is determined that there is a business plan that has not been selected as an evaluation target (Yes in S200), the processor 11 adds 1 to the value k (S210).

After that, the processor 11 newly selects the k-th business plan as an evaluation target based on the value k after addition (S160), and executes the processes from S170. In this way, the processor 11 selects a business plan to be evaluated in descending order of predicted results, and if the plan tolerance Z of the corresponding business plan exceeds a predetermined value, the corresponding business plan A proposal is made to members (S230).

On the other hand, if there is no business plan in which the plan tolerance Z exceeds a predetermined value among the plurality of business plans (No in S200), the processor 11 selects a business plan that proposes a specific business plan among the plurality of business plans. Decision is made (S220), and plan proposal data of the corresponding business plan is output (S230).

A particular business plan may be the business plan with the greatest planning tolerance among the plurality of business plans. Alternatively, among a plurality of business plans, the business plan with the largest expected result may be the business plan.

According to the information processing system 1 of the present embodiment described above, the processor 11 creates a plurality of business plans as candidates for the business plan to be proposed to the salesperson, using the predicted result as an index (S140). Then, the plan tolerance Z of the business plan is calculated in descending order of predicted results, and only the business plan with the plan tolerance Z exceeding a predetermined value is proposed to the salesperson.

Specifically, the processor 11 calculates the acceptance probability p _appr as an estimated value of the likelihood that the salesperson will approve the execution of the corresponding sales activities for each sales activity included in the sales plan to be evaluated. Calculate (S330).

The processor 11 collects the acceptance probabilities p _appr for each business activity, calculates the plan acceptance rate P _appr that is the degree of acceptance of the business plan (S350), and uses the plan acceptance rate P _appr to calculate the plan acceptance rate. Based on the degree Z, it is determined whether or not to propose the sales plan to the salesperson (S180). When the processor 11 determines to propose (Yes at S180), the processor 11 outputs plan proposal data for proposing the business plan to the salesperson (S230).

Therefore, according to the information processing system 1 of the present embodiment, there is no need to propose a sales plan with a low degree of tolerance to the sales staff, and the possibility that the sales activities based on the plan will be rejected by the sales staff can be considered. Can propose an appropriate business plan to employees.

In particular, in this embodiment, a plurality of different business plans are created (S140), and among the plurality of business plans, the business plan whose tolerance exceeds a predetermined value and which has the maximum expected result, The sales plan to be proposed to the salesperson is selected (S190).

The processor 11 calculates the plan acceptance rate _Pappr and the plan tolerance Z for the first business plan (S170). The operation (S170) of calculating the plan acceptance rate P _appr and the plan tolerance Z for the second business plan of the point is repeatedly executed, and as soon as a business plan whose plan tolerance Z exceeds a predetermined value is found, the business plan is put into business. The sales plan to be proposed to the employee is determined (S190).

Therefore, according to the information processing system 1 of the present embodiment, it is possible to propose to the salesperson a sales plan that is likely to be accepted by the salesperson, among the sales plans with high predicted results.

Furthermore, in this embodiment, the editing probabilities p _com1 and p _com2 for each change pattern are calculated as estimated values regarding the possibility of the salesperson changing the business plan (S430). Whether or not to propose the sales plan to the salesperson is determined based on the plan tolerance Z that takes into account the plan editing cost C _E derived from the editing probabilities p _com1 and p _com2 . The plan compilation cost _CE corresponds to an estimate of the change load of the business plan.

Thus, in the information processing system 1 of the present embodiment, it is determined whether or not to propose a business plan, taking into account the possibility of change of the plan and the change load predicted based thereon. Therefore, the information processing system 1 can propose an appropriate business plan to the salesperson with higher accuracy.

Furthermore, the processor 11 constructs an edit prediction model, which is a prediction model regarding the possibility of change of the plan, based on the history of changes to the business plans proposed in the past (S135). By constructing an edit prediction model based on the change history and calculating the edit probabilities p _com1 and p _com2 , it is possible to estimate the change possibility of the plan with high accuracy.

More specifically, processor 11 builds an acceptance prediction model based on the history of acceptance and rejection of each business activity included in previously proposed business plans. By constructing an acceptance prediction model based on the history of acceptance and refusal and calculating the acceptance probability _pappr , it is possible to estimate the possibility of acceptance of each business activity in the business plan with high accuracy.

[Other embodiments]
The present disclosure is not limited to the above embodiments, and can take various forms. For example, in the above embodiment, the sales plan to be proposed to the salesperson is selected using the plan acceptance rate Z based on the plan acceptance rate P _appr and the plan editing cost _CE as an index. However, the information processing system 1 may be configured to select a sales plan to be proposed to a salesperson using only one of the plan acceptance rate _Pappr and the plan editing cost _CE as an index.

Acceptance prediction models and edit prediction models may be constructed using machine learning techniques other than neural networks. For example, acceptance prediction models and editorial prediction models may be constructed using logistic regression, probabilistic models, or decision trees.

A function possessed by one component in the above embodiment may be distributed to a plurality of components. Functions possessed by multiple components may be integrated into one component. A part of the configuration of the above embodiment may be omitted. At least part of the configurations of the above embodiments may be added or replaced with respect to the configurations of other above embodiments. All aspects included in the technical ideas specified from the language in the claims are embodiments of the present disclosure.

[Technical concept disclosed in this specification]
It can be understood that the following technical ideas are disclosed in this specification.
[Item 1]
An information processing system for proposing a plan for sales activities,
a creation unit configured to create a business plan including one or more business activities, the business plan specifying the type and execution date and time of each of the one or more business activities;
a calculation unit configured to calculate, for each of the one or more sales activities included in the sales plan, an acceptance estimate, which is an estimate of the likelihood of a salesperson agreeing to perform the corresponding sales activity;
a determination unit configured to determine whether to propose the sales plan to the salesperson based on the acceptance estimate;
a proposal unit configured to output information for proposing the business plan to the salesperson when the determination unit determines to propose;
An information processing system comprising
[Item 2]
The creation unit creates a first business plan as the business plan, further creates a second business plan different from the first business plan,
The calculation unit calculates, as the acceptance estimate, an estimated value relating to the possibility that the salesperson will approve the execution of the corresponding sales activity for each of the one or more sales activities included in the first business plan. ,
The determination unit determines whether or not to propose the first business plan to the salesperson based on the estimated acceptance value related to the first business plan,
The proposal unit outputs information for proposing the first business plan when the determination unit determines to propose, and outputs the information for proposing the first business plan when the determination unit determines not to propose the second business plan. The information processing system according to item 1, which outputs information for proposing a plan to the salesperson.
[Item 3]
The calculation unit calculates the acceptance estimated value related to the first business plan as a first acceptance estimated value, and furthermore, with respect to each of the one or more business activities included in the second business plan, calculating, as a second acceptance estimate, an estimate of the likelihood that the salesperson will agree to perform the sales activity,
The proposal unit determines whether or not to propose the second business plan to the salesperson based on the second acceptance estimate when the determination unit determines not to propose, and makes a proposal. 3. The information processing system according to item 2, which outputs information for proposing the second business plan to the salesperson when it is determined that the second business plan is to be made.
[Item 4]
The calculation unit further calculates a change estimated value, which is an estimated value related to the possibility of the salesperson changing the business plan, with respect to a plurality of change patterns,
2. The information processing system according to item 1, wherein the determination unit determines whether or not to propose the sales plan to the salesperson based on the estimated acceptance value and the estimated change value.
[Item 5]
The determination unit calculates an estimated value regarding the change load of the business plan by the salesperson based on the change estimated value, and determines the business plan based on the acceptance estimated value and the estimated value regarding the change load of the business plan. 5. The information processing system according to item 4, which determines whether or not to make a proposal to the salesperson.
[Item 6]
The creation unit creates a first business plan as the business plan, further creates a second business plan different from the first business plan,
The calculation unit
with respect to each of said one or more sales activities included in said first sales plan, said acceptance estimate being an estimate relating to the likelihood of said salesperson agreeing to carry out the corresponding sales activity as a first acceptance estimate; and further calculating the plurality of change patterns with the change estimate value, which is an estimate value related to the possibility of the salesperson changing the first business plan, as the first change estimate value,
calculating, for each of the one or more sales activities included in the second sales plan, a second acceptance estimate that is an estimate of the likelihood that the salesperson will consent to perform the corresponding sales activity; and , calculating a second change estimate value, which is an estimate value related to the possibility of the salesperson changing the second sales plan, with respect to the plurality of change patterns;
The determination unit determines whether or not to propose the first business plan to the salesperson based on the first acceptance estimate and the first change estimate,
The proposal unit determines whether to propose the second sales plan to the salesperson based on the second acceptance estimate and the second change estimate when the determination unit determines not to propose. 6. The information processing system according to item 4 or 5, configured to propose the second business plan to the salesperson when it is determined to propose.
[Item 7]
The proposal unit calculates an estimated value related to the change load of the second business plan by the salesperson based on the second estimated value of change, and calculates the second estimated value of acceptance and the second business plan. 7. The information processing system according to item 6, wherein it is determined whether or not to propose the second business plan to the salesperson based on an estimated value of change load.
[Item 8]
a prediction model that outputs estimated values regarding the possibility of the salesperson changing the business plan with respect to the plurality of change patterns, based on the input of the type and execution date of at least one business activity included in the business plan; further comprising a model building unit configured to build based on a history of changes to previously proposed business plans;
6. The information processing system according to item 4 or 5, wherein the calculation unit calculates the change estimated value based on the prediction model.
[Item 9]
A previously proposed sales plan includes a predictive model that outputs an estimate of the likelihood that the salesperson will accept the at least one sales activity based on the input of the type of at least one sales activity and the execution date and time. further comprising a model builder configured to build based on a history of acceptance and rejection of each business activity performed;
The information processing system according to any one of items 1 to 7, wherein the calculation unit calculates the acceptance estimated value based on the prediction model.
[Item 10]
A computer program for causing a computer to implement the functions of the creating unit, the calculating unit, the determining unit, and the proposing unit in the information processing system according to any one of items 1 to 7.
[Item 11]
A computer-implemented information processing method comprising:
preparing a business plan including one or more business activities, the business plan specifying the type and execution date and time of each of the one or more business activities;
For each of the one or more sales activities included in the sales plan, calculating an acceptance estimate, which is an estimate of the likelihood of a salesperson agreeing to perform the corresponding sales activity;
determining whether to propose the sales plan to the salesperson based on the acceptance estimate;
outputting information for proposing the business plan to the salesperson when it is determined to propose;
Information processing method including.

DESCRIPTION OF SYMBOLS 1... Information processing system, 11... Processor, 12... Memory, 13... Storage, 19... Communication interface, 30... Information terminal.

Claims (12)

An information processing system for proposing a plan for sales activities,
a creation unit configured to create a business plan including one or more business activities, the business plan specifying the type and execution date and time of each of the one or more business activities;
For each of the one or more sales activities included in the sales plan, an acceptance estimate, which is an estimate of the likelihood that the salesperson will accept the corresponding sales activity , is calculated based on an acceptance prediction model. a calculation unit configured to:
a determination unit configured to determine whether to propose the sales plan to the salesperson based on the acceptance estimate;
a proposal unit configured to output information for proposing the business plan to the salesperson when the determination unit determines to propose;
wherein the acceptance prediction model is a prediction model that outputs an estimated value regarding the likelihood that the salesperson will accept the at least one sales activity based on the input of at least one type of sales activity and execution date and time. An information processing system that is a prediction model constructed based on the history of acceptance and rejection of each business activity included in a business plan proposed in the past. The creation unit creates a first business plan as the business plan, further creates a second business plan different from the first business plan,
The calculation unit calculates, as the acceptance estimate, an estimated value relating to the possibility that the salesperson will approve the execution of the corresponding sales activity for each of the one or more sales activities included in the first business plan. ,
The determination unit determines whether or not to propose the first business plan to the salesperson based on the estimated acceptance value related to the first business plan,
The proposal unit outputs information for proposing the first business plan when the determination unit determines to propose, and outputs the information for proposing the first business plan when the determination unit determines not to propose the second business plan. 2. The information processing system according to claim 1, wherein information for proposing a plan to said salesperson is output. The calculation unit calculates the acceptance estimated value related to the first business plan as a first acceptance estimated value, and furthermore, with respect to each of the one or more business activities included in the second business plan, calculating, as a second acceptance estimate, an estimate of the likelihood that the salesperson will agree to perform the sales activity,
The proposal unit determines whether or not to propose the second sales plan to the salesperson based on the second acceptance estimate when the determination unit determines not to propose, and makes a proposal. 3. The information processing system according to claim 2, wherein information for proposing said second sales plan to said salesperson is output when it is determined that said salesperson will do so. The calculation unit further calculates a change estimated value, which is an estimated value related to the possibility of the salesperson changing the business plan, with respect to a plurality of change patterns,
2. The information processing system according to claim 1, wherein said judgment unit judges whether or not to propose said business plan to said salesperson based on said acceptance estimated value and said change estimated value. The determination unit calculates an estimated value regarding the change load of the business plan by the salesperson based on the change estimated value, and determines the business plan based on the acceptance estimated value and the estimated value regarding the change load of the business plan. 5. The information processing system according to claim 4, wherein it is determined whether or not to propose to said salesperson. The creation unit creates a first business plan as the business plan, further creates a second business plan different from the first business plan,
The calculation unit
with respect to each of said one or more sales activities included in said first sales plan, said acceptance estimate being an estimate relating to the likelihood of said salesperson agreeing to carry out the corresponding sales activity as a first acceptance estimate; and further calculating the plurality of change patterns with the change estimate value, which is an estimate value related to the possibility of the salesperson changing the first business plan, as the first change estimate value,
calculating, for each of the one or more sales activities included in the second sales plan, a second acceptance estimate that is an estimate of the likelihood that the salesperson will consent to perform the corresponding sales activity; and , calculating a second change estimate value, which is an estimate value related to the possibility of the salesperson changing the second sales plan, with respect to the plurality of change patterns;
The determination unit determines whether or not to propose the first business plan to the salesperson based on the first acceptance estimate and the first change estimate,
The proposal unit determines whether to propose the second sales plan to the salesperson based on the second acceptance estimate and the second change estimate when the determination unit determines not to propose. 5. The information processing system according to claim 4, wherein said second business plan is proposed to said salesperson when it is determined that said second business plan should be proposed. The proposal unit calculates an estimated value related to the change load of the second business plan by the salesperson based on the second estimated value of change, and calculates the second estimated value of acceptance and the second business plan. 7. The information processing system according to claim 6, wherein it is determined whether or not to propose said second business plan to said salesperson based on an estimated value of change load. The calculation unit is configured to calculate the change estimated value based on a change prediction model,
The change prediction model estimates the possibility of the salesperson changing the business plan based on the input of at least one type of business activity and execution date included in the business plan, with respect to the plurality of change patterns. 6. The information processing system according to claim 4 , wherein the predictive model to be output is a predictive model constructed based on a history of changes to business plans proposed in the past. Claims 1 to 7, further comprising : an acceptance prediction model building unit configured to build the acceptance prediction model based on a history of acceptance and rejection of each business activity included in a previously proposed business plan. The information processing system according to any one of . A change prediction model building unit configured to build the change prediction model based on a history of changes to business plans proposed in the past.
The information processing system of claim 8, further comprising: 8. A computer program for causing a computer to implement the functions of the creating unit, the calculating unit, the determining unit, and the proposing unit in the information processing system according to any one of claims 1 to 7. A computer-implemented information processing method comprising:
preparing a business plan including one or more business activities, the business plan specifying the type and execution date and time of each of the one or more business activities;
For each of the one or more sales activities included in the sales plan, calculating an acceptance estimate, which is an estimate of the likelihood that the salesperson will accept the execution of the corresponding sales activity , based on an acceptance prediction model. ,
determining whether to propose the sales plan to the salesperson based on the acceptance estimate;
outputting information for proposing the business plan to the salesperson when it is determined to propose;
wherein the acceptance prediction model is a prediction model that outputs an estimated value regarding the likelihood that the salesperson will accept the at least one sales activity based on inputs of at least one type of sales activity and execution date and time An information processing method that is a prediction model constructed based on the history of acceptance and rejection of each business activity included in a business plan proposed in the past.

Images