Automated conference item selection

1 . A method, comprising:

determining, by a server, that text associated with speech in a call includes one or more keywords;

determining, by the server, a tone of the speech based on a speed of the speech;

determining, by the server, an emotional state of a caller based on the one or more keywords and the tone of the speech, wherein determining the emotional state comprises determining a context of a keyword using a machine learning model that analyzes a neighboring word range of the keyword and selecting a model based on location data of the caller;

retrieving, by the server, one or more items from a data store based on the emotional state of the caller and on at least one of a context of the one or more keywords, a weight assigned to the one or more keywords, a caller attribute, social media data associated with the caller, or spending data associated with the caller;

generating a user interface (UI) output based on an experience level of an agent and the one or more items, wherein the UI output automatically displays a selected script for an agent having a low experience level and the UI output displays a list of the one or more items in a priority order for an agent having a high experience level, and wherein the UI output is automatically adapted in real-time based on a detected change in a subject of the call;

transmitting, by the server, the UI output to an agent device for display; and

monitoring the call for additional keywords to dynamically update the UI output.

2 . The method of claim 1 , wherein determining the tone comprises analyzing a volume of the speech and a pitch of the speech.

3 . The method of claim 1 , wherein the one or more keywords includes a respective weight based on location data of the caller.

4 . The method of claim 1 , wherein retrieving the one or more items from the data store further comprises filtering the items based on a relevance score determined by a probabilistic match between the one or more keywords and data associated with each item.

5 . The method of claim 1 , comprising:

identifying, by the server, a caller identity based on a telephone caller identification (ID).

6 . The method of claim 1 , wherein the server is further configured to retrieve, from the data store, a daily bulletin regarding a system state in a geographic region corresponding to location data of the caller.

7 . The method of claim 1 , comprising:

detecting, by the server, a change in the subject of the call based on a sequence of detected keywords; and

automatically updating the UI output to display a different set of items relevant to a new subject.

8 . A system, comprising:

an agent device; and

a server configured to:

determine that text associated with speech in a call includes one or more keywords;

determine a tone of the speech based on a speed of the speech;

determine an emotional state of a caller based on the one or more keywords and the tone of the speech, wherein the server is configured to determine a context of a keyword using a machine learning model that analyzes a neighboring word range of the keyword and select a model based on location data of the caller;

retrieve one or more items from a data store based on the emotional state of the caller and on at least one of a context of the one or more keywords, a weight assigned to the one or more keywords, a caller attribute, social media data associated with the caller, or spending data associated with the caller;

generate a user interface (UI) output based on an experience level of an agent and the one or more items, wherein the UI output automatically displays a selected script for an agent having a low experience level and the UI output displays a list of the one or more items in a priority order for an agent having a high experience level, and wherein the UI output is automatically adapted in real-time based on a detected change in a subject of the call;

transmit the UI output to the agent device for display; and

monitor the call for additional keywords to dynamically update the UI output.

9 . The system of claim 8 , wherein the server is configured to:

detect an accent of the caller and refine the emotional state using location data.

10 . The system of claim 8 , wherein the one or more keywords includes a respective weight.

11 . The system of claim 8 , wherein the one or more keywords includes a respective weight based on location data of a caller.

12 . The system of claim 8 , wherein a number of social media followers of the caller is correlated with a social media presence strength.

13 . The system of claim 8 , wherein the server is configured to:

obtain location data of the caller; and

retrieve one or more items from the data store based on the location data of the caller.

14 . The system of claim 8 , wherein the server is configured to:

determine a context of the one or more keywords; and

retrieve one or more items from the data store based on the context of the one or more keywords.

15 . A non-transitory computer-readable medium comprising instructions stored on a memory, that when executed by a processor, cause the processor to perform operations comprising:

determining that text associated with speech in a call includes one or more keywords;

determining a tone of the speech based on a speed of the speech;

determining an emotional state of a caller based on the one or more keywords and the tone of the speech, wherein determining the emotional state comprises determining a context of a keyword using a machine learning model that analyzes a neighboring word range of the keyword and selecting a model based on location data of the caller;

retrieving one or more items from a data store based on the emotional state of the caller and on at least one of a context of the one or more keywords, a weight assigned to the one or more keywords, a caller attribute, social media data associated with the caller, or spending data associated with the caller;

generating a user interface (UI) output based on an experience level of an agent and the one or more items, wherein the UI output automatically displays a selected script for an agent having a low experience level and the UI output displays a list of the one or more items in a priority order for an agent having a high experience level, and wherein the UI output is automatically adapted in real-time based on a detected change in a subject of the call;

transmitting the UI output to an agent device for display; and

monitoring the call for additional keywords to dynamically update the UI output.

16 . The non-transitory computer-readable medium of claim 15 , wherein determining the tone comprises analyzing a volume of the speech and a pitch of the speech.

17 . The non-transitory computer-readable medium of claim 15 , wherein retrieving the one or more items from the data store further comprises filtering the items based on a relevance score determined by a probabilistic match between the one or more keywords and data associated with each item.

18 . The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise retrieving, from the data store, a daily bulletin regarding a system state in a geographic region corresponding to location data of the caller.

19 . The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise:

detecting a change in the subject of the call based on a sequence of detected keywords; and

automatically updating the UI output to display a different set of items relevant to a new subject.

20 . The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise:

identifying a caller identity based on a telephone caller identification (ID).