Dynamic field value recommendation methods and systems

Patent metadata

Field	Value
Publication number	US-11314790-B2
Application number	US-202015929364-A
Country	US
Kind code	B2
Filing date	Apr 28, 2020
Priority date	Nov 18, 2019
Publication date	Apr 26, 2022
Grant date	Apr 26, 2022

Abstract

Computing systems, database systems, and related methods are provided for recommending values for fields of database objects and dynamically updating a recommended value for a field of a database record in response to updated auxiliary data associated with the database record. One method involves obtaining associated conversational data, segmenting the conversational data, converting each respective segment of conversational data into a numerical representation, generating a combined numerical representation of the conversational data based on the sequence of numerical representations using an aggregation model, generating the recommended value based on the combined numerical representation of the conversational data using a prediction model associated with the field, and autopopulating the field of the case database object with the recommended value.

First claim

What is claimed is:

1. A method of providing a recommended value for an unstructured field of a database object, the method comprising:
obtaining correspondence associated with the database object, wherein the correspondence comprises a plurality of utterances by different speakers;
segmenting the correspondence into a plurality of conversational segments, wherein at least one conversational segment of the plurality of conversational segments includes textual content comprising an utterance of the plurality of utterances by a respective speaker of the different speakers;
for one or more of the plurality of conversational segments, encoding the textual content of the respective conversational segment comprising the utterance by the respective speaker into an encoded value representing the overall content of the respective conversational segment as a function of the textual content by inputting a text string corresponding to the respective conversational segment into an encoder model configured to generate the encoded value for the respective conversational segment as a function of the text string, resulting in a sequence of encoded values corresponding to the plurality of conversational segments of the correspondence;
generating a combined numerical representation of the correspondence based on the sequence of encoded values by inputting the sequence of encoded values to an aggregation model;
generating the recommended value for the unstructured field of the database object based on the combined numerical representation of the correspondence using a prediction model associated with the unstructured field, wherein the recommended value comprises a recommended summarization of the correspondence; and
autopopulating the unstructured field of the database object with the recommended value.

2. The method of claim 1 , wherein segmenting the correspondence comprises dividing chat messaging data on a per-utterance basis and each conversational segment of the plurality of conversational segments corresponds to a respective utterance of the chat messaging data.

3. The method of claim 1 , wherein the unstructured field comprises one of a summary field, a status field, or another field summarizing activity during a lifecycle of the database object.

4. The method of claim 1 , wherein the correspondence includes at least one of chat messaging data, a feed, and an e-mail thread.

5. The method of claim 1 , further comprising determining the aggregation model using historical values for the unstructured field of a reference set of existing database objects.

6. The method of claim 5 , wherein generating the recommended value comprises inputting the combined numerical representation into a summarization model configured to generate the recommended value as a function of the combined numerical representation, wherein the summarization model is derived using the historical values for the unstructured field of the reference set of existing database objects.

7. The method of claim 1 , wherein generating the recommended value comprises inputting the combined numerical representation into a summarization model configured to generate the recommended value as a function of the combined numerical representation.

8. The method of claim 1 , wherein generating the recommended value comprises inputting the combined numerical representation into an extractive summarization model configured to generate the recommended value as a function of the combined numerical representation by identifying and selecting a subset of the correspondence associated with the database object that is correlative to an extracted topic identified using the combined numerical representation.

9. The method of claim 1 , wherein generating the recommended value comprises inputting the combined numerical representation into an abstractive summarization model configured to generate the recommended value as a function of the combined numerical representation based on historical relationships between historical values for the unstructured field of a reference set of existing database objects and corresponding combined numerical representations of the reference set of existing database objects.

10. A computing system comprising:
a database to maintain a database object; and
a server coupled to the database and a network to obtain correspondence associated with the database object, wherein the correspondence comprises a plurality of utterances by different speakers, segment the correspondence into a plurality of conversational segments prior to determining a numerical sequence of encoded values corresponding to the plurality of conversational segments using an encoder model as a function of textual content of the plurality of conversational segments, wherein at least one conversational segment of the plurality of conversational segments includes textual content comprising an utterance of the plurality of utterances by a respective speaker of the different speakers and determining the numerical sequence of encoded values comprises inputting a text string corresponding to a respective conversational segment into the encoder model configured to generate the encoded value for the respective conversational segment as a function of the text string, convert the numerical sequence of encoded values into a combined numerical representation using an aggregation model by inputting the sequence of encoded values to the aggregation model, and determine a recommended value for a field of the database object based on the combined numerical representation, wherein the recommended value comprises a recommended summarization of the correspondence, and set the field of the database object to the recommended value.

11. The computing system of claim 10 , wherein the server generates a graphical indication of the recommended value for the field within a graphical user interface display associated with an instance of an application provided to a client device over the network.

12. The computing system of claim 10 , wherein:
the correspondence includes at least one of chat messaging data, a feed, and an e-mail thread;
the field comprises an unstructured field; and
the recommended value comprises an autogenerated summary of the correspondence determined as a function of the combined numerical representation using a summarization model.

13. The computing system of claim 10 , wherein:
the correspondence includes at least one of chat messaging data, a feed, and an e-mail thread;
the field comprises a structured field; and
the recommended value comprises an identified value from among a plurality of potential values for the structured field determined based on the combined numerical representation using a classification model.

14. A non-transitory computer-readable medium having instructions stored thereon that, when executed by a processing system, cause the processing system to:
obtain correspondence associated with a database object, wherein the correspondence comprises a plurality of utterances by different speakers;
segment the correspondence into a plurality of conversational segments, wherein at least one conversational segment of the plurality of conversational segments includes textual content comprising an utterance of the plurality of utterances by a respective speaker of the different speakers;
for one or more of the plurality of conversational segments, encode the textual content of the respective conversational segment comprising the utterance by the respective speaker into an encoded value representing the overall content of the respective conversational segment as a function of the textual content by inputting a

Assignees

• Salesforce Com Inc

Inventors

• Chang Son Thanh
• Peng Weiping
• Cheng Na
• Jiang Feifei
• Huffman Jacob Nathaniel
• Kumar Nandini Suresh
• Le Khoa
• Larry christopher

Classifications

• H04L51/56

  Unified messaging, e.g. interactions between e-mail, instant messaging or converged IP messaging [CPM] · CPC title

• G06F16/345

  Summarisation for human users · CPC title

• G06F16/24564

  Applying rules; Deductive queries · CPC title

• G06F16/328 Primary

  Management therefor · CPC title

• G06F16/24578

  using ranking · CPC title