Determining topics and action items from conversations

1. A method for organizing conversation information over a network using one or more network computers that include one or more processors to perform actions, comprising:

employing two or more machine learning (ML) models and a plurality of sentences provided from a conversation to generate an insight score for each of the plurality of sentences, wherein each insight score is based on two or more partial insight scores that each correspond to an ML model, and wherein each insight score correlates to a probability that its sentence includes one or more of an action or a question;

in response to one or more of the plurality of sentences being associated with one or more insight scores that exceed a threshold value, performing further actions, including:

determining an information score for each of the one or more of the plurality of sentences based on a number of included object words, a number of included subject words, a number of included insight words, and one or more included grammatical features;

determining a definiteness score for each of the one or more of the plurality of sentences based on decision-making words included in the one or more of the plurality of sentences, wherein each sentence that includes an amount of words associated with conclusive decision-making words are scored higher than each other sentence that includes another amount of words associated with conclusive decision-making words that is less than the amount of words associated with conclusive decision-making words;

generating one or more insights associated with the conversation based on the one or more of the plurality of sentences, the determined information score for each of the one or more of the plurality of sentences, and the determined definiteness score for each of the one or more of the plurality of sentences; and

generating a report that associates the one or more insights with one or more portions of the conversation that include the one or more of the plurality of sentences and that are associated with the one or more insights.

2. The method of claim 1 , further comprising:

determining the two or more ML models based on one or more of locale information, subject matter of the conversation, or user preferences, wherein each ML model is paired with an adversarial ML model.

3. The method of claim 1 , further comprising:

determining one or more dominant portions of the conversation based on a conversation digest that corresponds to the conversation;

determining one or more key subjects that respectively correspond to each of the one or more dominant portions of the conversation, wherein each of the one or more key subjects is a dominant subject in a respective dominant portion of the conversation; and

generating one or more topics of the conversation based on the one or more key subjects.

4. The method of claim 1 , wherein generating the one or more insights further comprises:

determining one or more action insights that include one or more of an open action item, an unclaimed action item, a closed action item, a follow-up, a task, an idea, or an appointment; and

determining one or more question insights that include one or more of an unanswered question, an answered question, or a request.

5. The method of claim 1 , wherein generating the one or more insights further comprises:

determining one or more insight models based on the conversation, wherein each insight model defines a set of one or more rules for generating the one or more insights;

executing each set of one or more rules to identify one or more words that are included in the one or more of the plurality of sentences and that are associated with one or more actions or one or more questions;

modifying the one or more identified words by one or more of removing a superfluous word, correcting grammar, or correcting punctuation; and

generating the one or more insights based on the modified one or more identified words.

6. The method of claim 1 , wherein generating the report further comprises:

generating, based on the one or more insights, one or more of a task, an appointment, a reminder, or a notification;

providing the one or more of the task, the appointment, the reminder, or the notification to one or more external calendar services.

7. The method of claim 1 , wherein generating the report further comprises:

generating a user interface that displays an interactive outline of the conversation, wherein one or more descriptions or one or more labels associated with the one or more insights are displayed as one or more outline headings; and

in response to a user selecting one or more of the one or more outline headings, displaying one or more portions of the conversation that are associated with one or more insights that are associated with the selected one or more outline headings.

8. A system for organizing conversation information over a network, comprising:

a network computer, comprising:

a memory that stores at least instructions; and

one or more processors that are configured to execute instructions to perform actions, wherein the actions include:

employing two or more machine learning (ML) models and a plurality of sentences provided from a conversation to generate an insight score for each of the plurality of sentences, wherein each insight score is based on two or more partial insight scores that each correspond to an ML model, and wherein each insight score correlates to a probability that its sentence includes one or more of an action or a question;

in response to one or more of the plurality of sentences being associated with one or more insight scores that exceed a threshold value, performing further actions, including:

determining an information score for each of the one or more of the plurality of sentences based on a number of included object words, a number of included subject words, a number of included insight words, and one or more included grammatical features;

determining a definiteness score for each of the one or more of the plurality of sentences based on decision-making words included in the one or more of the plurality of sentences, wherein each sentence that includes an amount of words associated with conclusive decision-making words are scored higher than each other sentence that includes another amount of words associated with conclusive decision-making words that is less than the amount of words associated with conclusive decision-making words;

generating one or more insights associated with the conversation based on the one or more of the plurality of sentences, the determined information score for each of the one or more of the plurality of sentences, and the determined definiteness score for each of the one or more of the plurality of sentences; and

generating a report that associates the one or more insights with one or more portions of the conversation that include the one or more of the plurality of sentences and that are associated with the one or more insights; and

a client computer, comprising:

another memory that stores at least instructions; and

one or more other processors that are configured to execute instructions to perform other actions, wherein the other actions include:

providing the plurality of sentences to the network computer; and

displaying the report.

9. The system of claim 8 , wherein the one or more processors of the network computer are further configured to execute instructions to perform at least one additional action, wherein the at least one additional action includes:

determining the two or more ML models based on one or more of locale information, subject matter of the conversation, or user preferences, wherein each ML model is paired with an adversarial ML model.

10. The system of claim 8 , wherein the one or more processors of the network computer are further configured to execute instructions to perform additional actions, wherein the additional actions include:

determining one or more dominant portions of the conversation based on a conversation digest that corresponds to the conversation;

determining one or more key subjects that respectively correspond to each of the one or more dominant portions of the conversation, wherein each of the one or more key subjects is a dominant subject in a respective dominant portion of the conversation; and

generating one or more topics of the conversation based on the one or more key subjects.

11. The system of claim 8 , wherein generating the one or more insights further comprises:

determining one or more action insights that include one or more of an open action item, an unclaimed action item, a closed action item, a follow-up, a task, an idea, or an appointment; and

determining one or more question insights that include one or more of an unanswered question, an answered question, or a request.

12. The system of claim 8 , wherein generating the one or more insights further comprises:

determining one or more insight models based on the conversation, wherein each insight model defines a set of one or more rules for generating the one or more insights;

executing each set of one or more rules to identify one or more words that are included in the one or more of the plurality of sentences and that are associated with one or more actions or one or more questions;

modifying the one or more identified words by one or more of removing a superfluous word, correcting grammar, or correcting punctuation; and

generating the one or more insights based on the modified one or more identified words.

13. The system of claim 8 , wherein generating the report further comprises:

generating, based on the one or more insights, one or more of a task, an appointment, a reminder, or a notification;

providing the one or more of the task, the appointment, the reminder, or the notification to one or more external calendar services.

14. The system of claim 8 , wherein generating the report further comprises:

generating a user interface that displays an interactive outline of the conversation, wherein one or more descriptions or one or more labels associated with the one or more insights are displayed as one or more outline headings; and

in response to a user selecting one or more of the one or more outline headings, displaying one or more portions of the conversation that are associated with one or more insights that are associated with the selected one or more outline headings.

15. A processor readable non-transitory storage medium that includes instructions for organizing conversation information over a network, wherein execution of the instructions by one or more hardware processors causes actions to be performed, wherein the actions comprise:

employing two or more machine learning (ML) models and a plurality of sentences provided from a conversation to generate an insight score for each of the plurality of sentences, wherein each insight score is based on two or more partial insight scores that each correspond to an ML model, and wherein each insight score correlates to a probability that its sentence includes one or more of an action or a question;

in response to one or more of the plurality of sentences being associated with one or more insight scores that exceed a threshold value, performing further actions, including:

determining an information score for each of the one or more of the plurality of sentences based on a number of included object words, a number of included subject words, a number of included insight words, and one or more included grammatical features;

determining a definiteness score for each of the one or more of the plurality of sentences based on decision-making words included in the one or more of the plurality of sentences, wherein each sentence that includes an amount of words associated with conclusive decision-making words are scored higher than each other sentence that includes another amount of words associated with conclusive decision-making words that is less than the amount of words associated with conclusive decision-making words;

generating one or more insights associated with the conversation based on the one or more of the plurality of sentences, the determined information score for each of the one or more of the plurality of sentences, and the determined definiteness score for each of the one or more of the plurality of sentences; and

generating a report that associates the one or more insights with one or more portions of the conversation that include the one or more of the plurality of sentences and that are associated with the one or more insights.

16. The medium of claim 15 , wherein execution of the instructions by the one or more hardware processors causes at least one additional action to be performed, wherein the at least one additional action includes:

determining the two or more ML models based on one or more of locale information, subject matter of the conversation, or user preferences, wherein each ML model is paired with an adversarial ML model.

17. The medium of claim 15 , wherein execution of the instructions by the one or more hardware processors causes additional actions to be performed, wherein the additional actions include:

determining one or more dominant portions of the conversation based on a conversation digest that corresponds to the conversation;

determining one or more key subjects that respectively correspond to each of the one or more dominant portions of the conversation, wherein each of the one or more key subjects is a dominant subject in a respective dominant portion of the conversation; and

generating one or more topics of the conversation based on the one or more key subjects.

18. The medium of claim 15 , wherein generating the one or more insights further comprises:

determining one or more action insights that include one or more of an open action item, an unclaimed action item, a closed action item, a follow-up, a task, an idea, or an appointment; and

determining one or more question insights that include one or more of an unanswered question, an answered question, or a request.

19. The medium of claim 15 , wherein generating the one or more insights further comprises:

determining one or more insight models based on the conversation, wherein each insight model defines a set of one or more rules for generating the one or more insights;

executing each set of one or more rules to identify one or more words that are included in the one or more of the plurality of sentences and that are associated with one or more actions or one or more questions;

modifying the one or more identified words by one or more of removing a superfluous word, correcting grammar, or correcting punctuation; and

generating the one or more insights based on the modified one or more identified words.

20. The medium of claim 15 , wherein generating the report further comprises:

generating, based on the one or more insights, one or more of a task, an appointment, a reminder, or a notification;

providing the one or more of the task, the appointment, the reminder, or the notification to one or more external calendar services.

21. The medium of claim 15 , wherein generating the report further comprises:

generating a user interface that displays an interactive outline of the conversation, wherein one or more descriptions or one or more labels associated with the one or more insights are displayed as one or more outline headings; and

in response to a user selecting one or more of the one or more outline headings, displaying one or more portions of the conversation that are associated with one or more insights that are associated with the selected one or more outline headings.

22. A network computer for organizing conversation information over a network, comprising:

a memory that stores at least instructions; and

one or more processors that are configured to execute instructions to perform actions, wherein the actions include:

employing two or more machine learning (ML) models and a plurality of sentences provided from a conversation to generate an insight score for each of the plurality of sentences, wherein each insight score is based on two or more partial insight scores that each correspond to an ML model, and wherein each insight score correlates to a probability that its sentence includes one or more of an action or a question;

in response to one or more of the plurality of sentences being associated with one or more insight scores that exceed a threshold value, performing further actions, including:

determining an information score for each of the one or more of the plurality of sentences based on a number of included object words, a number of included subject words, a number of included insight words, and one or more included grammatical features;

determining a definiteness score for each of the one or more of the plurality of sentences based on decision-making words included in the one or more of the plurality of sentences, wherein each sentence that includes an amount of words associated with conclusive decision-making words are scored higher than each other sentence that includes another amount of words associated with conclusive decision-making words that is less than the amount of words associated with conclusive decision-making words;

generating one or more insights associated with the conversation based on the one or more of the plurality of sentences, the determined information score for each of the one or more of the plurality of sentences, and the determined definiteness score for each of the one or more of the plurality of sentences; and

generating a report that associates the one or more insights with one or more portions of the conversation that include the one or more of the plurality of sentences and that are associated with the one or more insights.

23. The network computer of claim 22 , wherein the one or more processors execute the instructions to perform at least one additional action, wherein the at least one additional action includes:

determining the two or more ML models based on one or more of locale information, subject matter of the conversation, or user preferences, wherein each ML model is paired with an adversarial ML model.

24. The network computer of claim 22 , wherein the one or more processors execute the instructions to perform additional actions, wherein the additional actions include:

determining one or more dominant portions of the conversation based on a conversation digest that corresponds to the conversation;

determining one or more key subjects that respectively correspond to each of the one or more dominant portions of the conversation, wherein each of the one or more key subjects is a dominant subject in a respective dominant portion of the conversation; and

generating one or more topics of the conversation based on the one or more key subjects.

25. The network computer of claim 22 , wherein generating the one or more insights further comprises:

determining one or more action insights that include one or more of an open action item, an unclaimed action item, a closed action item, a follow-up, a task, an idea, or an appointment; and

determining one or more question insights that include one or more of an unanswered question, an answered question, or a request.

26. The network computer of claim 22 , wherein generating the one or more insights further comprises:

determining one or more insight models based on the conversation, wherein each insight model defines a set of one or more rules for generating the one or more insights;

executing each set of one or more rules to identify one or more words that are included in the one or more of the plurality of sentences and that are associated with one or more actions or one or more questions;

modifying the one or more identified words by one or more of removing a superfluous word, correcting grammar, or correcting punctuation; and

generating the one or more insights based on the modified one or more identified words.

27. The network computer of claim 22 , wherein generating the report further comprises:

generating, based on the one or more insights, one or more of a task, an appointment, a reminder, or a notification;

providing the one or more of the task, the appointment, the reminder, or the notification to one or more external calendar services.

28. The network computer of claim 22 , wherein generating the report further comprises:

generating a user interface that displays an interactive outline of the conversation, wherein one or more descriptions or one or more labels associated with the one or more insights are displayed as one or more outline headings; and

in response to a user selecting one or more of the one or more outline headings, displaying one or more portions of the conversation that are associated with one or more insights that are associated with the selected one or more outline headings.