System and method for automatic process error detection and correction

1. A method for detecting an error in a business process via an exchange of email messages, comprising:

receiving, by a processor of a business process analysis server (BPAS) that receives all email messages for analysis, an email, wherein the email includes an address of the BPAS that is automatically added to each email that is created at an endpoint device of a user and a recipient email address when the email is created at the endpoint device, wherein the address of the BPAS is included in a carbon copy field or a blind carbon copy field of the email;

analyzing, by the processor, the email to determine at least one feature, wherein the at least one feature comprises a keyword within text from the email and a role associated with each email address in the email found from a company directory;

determining, by the processor, the business process based on the role associated with each email address and the keyword within the text in the email using a logistic regression model, the logistic regression model comprising,

P

⁡

(

p

|

m

1

:

;

θ

)

=

e

θ

T

⁢

φ

⁡

(

p

,

m

1

:

)

Σ

p

′

∈

ℙ

⁢

e

θ

T

⁢

φ

⁡

(

p

,

m

1

:

)

,

where P is a business process that is identified based upon a given process p and a message thread prefix m 1: from a training data set of pre-defined business processes and associated features that are known, θ is a vector of learned parameters from a thread of the emails, T, used to train the logistic regression model and, φ is a fixed dimension feature vector function of the given process p and the message thread prefix m 1: ;

determining, by the processor, one or more variables that is associated with activities associated with the business process based on analysis of the text within the email, wherein the analysis deploys a conditional random field model for a word-label sequence of the email of an email thread;

detecting, by the processor, the error in the business process associated with the email based on at least one variable of the one or more variables associated with the business process, wherein the error comprises at least one of: a role anomaly, an information anomaly or a workflow anomaly, wherein the detecting comprises:

comparing, by the processor, the activities and values of the one or more variables of the activities of the business process to known activities and associated variables of the business process; and

determining, by the processor, that a value of the at least one variable of the one or more variables of the activities is different than a value of a corresponding variable of the associated variables of the known activities; and

generating, by the processor, an alert email in response to the error that is detected, wherein the alert email requests a correction to the at least one variable to complete the business process, wherein the alert email is generated by the BPAS and addressed only to a user that has a particular role that is responsible for providing correct information for the at least one variable that caused the error in the business process, wherein the user is missing from a thread of emails associated with the business process.

2. The method of claim 1 , wherein the at least one feature comprises a plurality of features.

3. The method of claim 2 , wherein the business process is determined based on a comparison of the plurality of features to a set of features associated with each one of a plurality pre-determined business processes stored in a database in communication with the BPAS.

4. The method of claim 1 , wherein the one or more variables comprises at least one of: a party, an activity, a duration or a progress.

5. The method of claim 1 , wherein the error comprises that the at least one variable is missing.

6. The method of claim 1 , wherein the error comprises that the at least one variable is incorrect.

7. A non-transitory computer-readable medium storing a plurality of instructions, which when executed by a processor of a business process analysis server (BPAS), causes the processor to perform operations for detecting an error in a business process via an exchange of email messages comprising:

receiving an email, wherein the email includes an address of the BPAS that is automatically added to each email that is created at an endpoint device of a user and a recipient email address, wherein the BPAS receives all email messages for analysis when the email is created at the endpoint device, wherein the address of the BPAS is included in a carbon copy field or a blind carbon copy field of the email;

analyzing the email to determine at least one feature, wherein the at least one feature comprises a keyword within text from the email and a role associated with each email address in the email found from a company directory;

determining the business process based on the role associated with each email address and the keyword within the text in the email using a logistic regression model, the logistic regression model comprising,

P

⁡

(

p

|

m

1

:

;

θ

)

=

e

θ

T

⁢

φ

⁡

(

p

,

m

1

:

)

Σ

p

′

∈

ℙ

⁢

e

θ

T

⁢

φ

⁡

(

p

,

m

1

:

)

,

where P is a business process that is identified based upon a given process p and a message thread prefix m 1: from a training data set of pre-defined business processes and associated features that are known, θ is a vector of learned parameters from a thread of the emails, T, used to train the logistic regression model and, φ is a fixed dimension feature vector function of the given process p and the message thread prefix m 1: ;

determining one or more variables that is associated with activities associated with the business process based on analysis of the text within the email, wherein the analysis deploys a conditional random field model for a word-label sequence of the email of an email thread;

detecting the error in the business process associated with the email based on at least one variable of the one or more variables associated with the business process, wherein the error comprises at least one of: a role anomaly, an information anomaly or a workflow anomaly, wherein the detecting comprises:

comparing the activities and values of the one or more variables of the activities of the business process to known activities and associated variables of the business process; and

determining that a value of the at least one variable of the one or more variables of the activities is different than a value of a corresponding variable of the associated variables of the known activities; and

generating an alert email in response to the error that is detected, wherein the alert email requests a correction to the at least one variable to complete the business process, wherein the alert email is generated by the BPAS and addressed only to a user that has a particular role that is responsible for providing correct information for the at least one variable that caused the error in the business process, wherein the user is missing from a thread of emails associated with the business process.

8. The non-transitory computer-readable medium of claim 7 , wherein the at least one feature comprises a plurality of features.

9. The non-transitory computer-readable medium of claim 8 , wherein the business process is determined based on a comparison of the plurality of features to a set of features associated with each one of a plurality pre-determined business processes stored in a database in communication with the BPAS.

10. The non-transitory computer-readable medium of claim 7 , wherein the one or more variables comprises at least one of: a party, an activity, a duration or a progress.

11. The non-transitory computer-readable medium of claim 7 , wherein the error comprises that the at least one variable is missing.

12. The non-transitory computer-readable medium of claim 7 , wherein the error comprises that the at least one variable is incorrect.