Abstract: Certain aspects of the present disclosure provide techniques for training and using a machine learning model to extract relevant textual content for custom fields in a software application from freeform text samples. An example method generally includes generating, via a natural language processing pipeline, a training data set from a data set of freeform text samples and field entries for a plurality of custom fields defined in a software application. A first machine learning model is trained to identify custom fields for which relevant data is included in freeform text. A second machine learning model is trained to extract content from the freeform text into one or more custom fields of the plurality of custom fields defined in the software application and identified by the first machine learning model as custom fields for which relevant data is included in the freeform text.

Abstract: Systems and methods for scoring potential actions are disclosed. An example method may be performed by one or more processors of a system and include training a machine learning model based at least in part on a sequential database and retention data, identifying an action subsequence executed by a user, generating, for each of a plurality of potential actions, using the machine learning model, a first value indicating a probability that the user will execute the potential action immediately after executing the action subsequence, a second value indicating a probability that the user will continue to use the system if the user executes the potential action immediately after executing the action subsequence, and a confidence score indicating a likelihood that recommending the potential action to the user will result in the user continuing to use the system, the confidence score generated based on the first value and the second value.

Abstract: Certain aspects of the present disclosure provide techniques for training and using a machine learning model to extract relevant textual content for custom fields in a software application from freeform text samples. An example method generally includes generating, via a natural language processing pipeline, a training data set from a data set of freeform text samples and field entries for a plurality of custom fields defined in a software application. A first machine learning model is trained to identify custom fields for which relevant data is included in freeform text. A second machine learning model is trained to extract content from the freeform text into one or more custom fields of the plurality of custom fields defined in the software application and identified by the first machine learning model as custom fields for which relevant data is included in the freeform text.

Abstract: A method predicts custom fields from text. Transaction text is normalized from transaction data to generate normalized text. A field prediction and a type prediction are selected using prediction data and the normalized text. The prediction data is generated using a machine learning model trained to identify field predictions from free form text. The field prediction and the type prediction are presented to a client device. In response to user input from the client device, the transaction data is updated with the field prediction.

Abstract: Aspects of the present disclosure relate to real-time invoice error prevention. Embodiments include receiving a value related to an item or service during creation of an invoice by a user via a user interface, and determining a user-level mean and a user-level standard deviation related to the value based on historical invoices of the user. Embodiments include determining a global mean and a global standard deviation related to the value based on historical invoices of a plurality of users. Embodiments include selecting weights for the user-level mean, the user-level standard deviation, the global mean, and the global standard deviation based on a total number of the historical invoices of the user. Embodiments include determining an expected range for the value based on the user-level mean, the user-level standard deviation, the global mean, the global standard deviation, and the weights. Embodiments include determining that the value is outside the expected range.

Abstract: Systems and methods for scoring potential actions are disclosed. An example method may be performed by one or more processors of a system and include training a machine learning model based at least in part on a sequential database and retention data, identifying an action subsequence executed by a user, generating, for each of a plurality of potential actions, using the machine learning model, a first value indicating a probability that the user will execute the potential action immediately after executing the action subsequence, a second value indicating a probability that the user will continue to use the system if the user executes the potential action immediately after executing the action subsequence, and a confidence score indicating a likelihood that recommending the potential action to the user will result in the user continuing to use the system, the confidence score generated based on the first value and the second value.

Abstract: A method predicts custom fields from text. Transaction text is normalized from transaction data to generate normalized text. A field prediction and a type prediction are selected using prediction data and the normalized text. The prediction data is generated using a machine learning model trained to identify field predictions from free form text. The field prediction and the type prediction are presented to a client device. In response to user input from the client device, the transaction data is updated with the field prediction.