Defining and executing custom generative artificial intelligence agent actions on selected data items
Techniques for creating and executing custom AI-driven actions in a data analysis interface are disclosed. One or more embodiments receive user input that defines a custom action by specifying an AI prompt with a natural language query and context variables. This definition is stored for later use. One or more embodiments present a graphical user interface (GUI) displaying data items, allowing users to select a subset. Based on the stored definition of a custom action, an option to execute the custom action appears. When the option is chosen, one or more embodiments assign metadata values from the selected items to the context variables, creating initialized context variables. The complete prompt, including the query and initialized variables, is submitted to an AI agent. The AI's response is displayed in the GUI, providing context-aware insights within the data analysis workflow.
1 . One or more non-transitory computer-readable media comprising instructions which, when executed by one or more hardware processors, cause performance of operations comprising:
presenting an action definition interface for defining a custom action;
receiving, via the action definition interface, first user input that defines a generative artificial intelligence (AI) agent prompt comprising: (a) a natural language query; and (b) two or more context variables that provide context for the natural language query;
based at least in part on the first user input, storing a definition of the custom action comprising the generative AI agent prompt;
presenting a graphical user interface (GUI) comprising a visual representation of a plurality of data items;
receiving, via the GUI, second user input that selects two or more data items of the plurality of data items;
based on the stored definition of the custom action: presenting, in the GUI, an option to execute the custom action for the two or more data items of the plurality of data items;
receiving, via the GUI, third user input selecting the option to execute the custom action for the two or more data items of the plurality of data items;
responsive to receiving the third user input:
assigning values of metadata associated with the two or more data items of the plurality of data items to the two or more context variables to yield two or more initialized context variables;
submitting the generative AI agent prompt, comprising the natural language query and the two or more initialized context variables, to a generative AI agent; and
presenting, in the GUI, a first answer received from the generative AI agent in response to the generative AI agent prompt.
2 . The one or more non-transitory computer-readable media of claim 1 :
wherein presenting the action definition interface is performed for a first user who has credentials to define the custom action; and
wherein receiving the second user input that selects the two or more data items of the plurality of data items is performed for a second user who does not have credentials to define the custom action.
3 . The one or more non-transitory computer-readable media of claim 1 :
wherein the two or more data items of plurality of data items correspond to a plurality of services deployed respectively to a plurality of nodes in a multi-node system,
wherein the plurality of services is associated with a log analytics system.
4 . The one or more non-transitory computer-readable media of claim 1 , wherein the first user input selects the generative AI agent from a plurality of endpoints, wherein at least one of the plurality of endpoints is not LLM-based.
5 . The one or more non-transitory computer-readable media of claim 1 , wherein the first user input selects a container for executing the generative AI agent prompt from a plurality of containers in a cloud environment.
6 . The one or more non-transitory computer-readable media of claim 1 , wherein the first user input defines one or more transformations to apply to the two or more initialized context variables before submitting the generative AI prompt to the generative AI agent.
7 . The one or more non-transitory computer-readable media of claim 1 , wherein the visual representation of the plurality of data items comprises a first chart that represents a first two or more data items of the plurality of data items and a second chart that represents a second two or more data items of the plurality of data items.
8 . The one or more non-transitory computer-readable media of claim 7 , wherein the second user input comprises selections from both the first chart and the second chart.
9 . The one or more non-transitory computer-readable media of claim 1 , the operations further comprising:
receiving fourth user input comprising a natural-language follow-up question;
responsive to receiving the fourth user input, submitting the natural-language follow-up question to the generative AI agent; and
presenting, in the GUI, a second answer received from the generative AI agent in response to the natural-language follow-up question.
10 . The one or more non-transitory computer-readable media of claim 9 , wherein submitting the natural-language follow-up question to the generative AI agent comprises resubmitting the two or more initialized context variables to the generative AI agent.
11 . The one or more non-transitory computer-readable media of claim 1 , the operations further comprising:
based on the natural language query:
generating a non-LLM query for two or more of the values of metadata associated with the two or more data items of the plurality of data items; and
submitting the non-LLM query to a non-LLM system obtain the two or more of the values of metadata associated with the two or more data items of the plurality of data items.
12 . The one or more non-transitory computer-readable media of claim 11 , the operations further comprising:
training a machine learning model to generate non-LLM queries for values of metadata based on natural language queries.
13 . A method comprising:
presenting an action definition interface for defining a custom action;
receiving, via the action definition interface, first user input that defines a generative artificial intelligence (AI) agent prompt comprising: (a) a natural language query; and (b) two or more context variables that provide context for the natural language query;
based at least in part on the first user input, storing a definition of the custom action comprising the generative AI agent prompt;
presenting a graphical user interface (GUI) comprising a visual representation of a plurality of data items;
receiving, via the GUI, second user input that selects two or more data items of the plurality of data items;
based on the stored definition of the custom action: presenting, in the GUI, an option to execute the custom action for the two or more data items of the plurality of data items;
receiving, via the GUI, third user input selecting the option to execute the custom action for the two or more data items of the plurality of data items;
responsive to receiving the third user input:
assigning values of metadata associated with the two or more data items of the plurality of data items to the two or more context variables to yield two or more initialized context variables;
submitting the generative AI agent prompt, comprising the natural language query and the two or more initialized context variables, to a generative AI agent; and
presenting, in the GUI, a first answer received from the generative AI agent in response to the generative AI agent prompt; and
wherein the method is performed by at least one device including a hardware processor.
14 . The method of claim 13 :
wherein presenting the action definition interface is performed for a first user who has credentials to define the custom action; and
wherein receiving the second user input that selects the two or more data items of the plurality of data items is performed for a second user who does not have credentials to define the custom action.
15 . The method of claim 13 :
wherein the two or more data items of plurality of data items correspond to a plurality of services deployed respectively to a plurality of nodes in a multi-node system,
wherein the plurality of services is associated with a log analytics system.
16 . The method of claim 13 , wherein the first user input selects the generative AI agent from a plurality of endpoints, wherein at least one of the plurality of endpoints is not LLM-based.
17 . A system comprising:
at least one device including a hardware processor;
the system being configured to perform operations comprising:
presenting an action definition interface for defining a custom action;
receiving, via the action definition interface, first user input that defines a generative artificial intelligence (AI) agent prompt comprising: (a) a natural language query; and (b) two or more context variables that provide context for the natural language query;
based at least in part on the first user input, storing a definition of the custom action comprising the generative AI agent prompt;
presenting a graphical user interface (GUI) comprising a visual representation of a plurality of data items;
receiving, via the GUI, second user input that selects two or more data items of the plurality of data items;
based on the stored definition of the custom action: presenting, in the GUI, an option to execute the custom action for the two or more data items of the plurality of data items;
receiving, via the GUI, third user input selecting the option to execute the custom action for the two or more data items of the plurality of data items;
responsive to receiving the third user input:
assigning values of metadata associated with the two or more data items of the plurality of data items to the two or more context variables to yield two or more initialized context variables;
submitting the generative AI agent prompt, comprising the natural language query and the two or more initialized context variables, to a generative AI agent; and
presenting, in the GUI, a first answer received from the generative AI agent in response to the generative AI agent prompt.
18 . The system of claim 17 , wherein the first user input defines one or more transformations to apply to the two or more initialized context variables before submitting the generative AI prompt to the generative AI agent.
19 . The system of claim 17 , wherein the visual representation of the plurality of data items comprises a first chart that represents a first two or more data items of the plurality of data items and a second chart that represents a second two or more data items of the plurality of data items; and wherein the second user input comprises selections from both the first chart and the second chart.
20 . The system of claim 17 , the operations further comprising:
receiving fourth user input comprising a natural-language follow-up question;
responsive to receiving the fourth user input, submitting the natural-language follow-up question to the generative AI agent;
presenting, in the GUI, a second answer received from the generative AI agent in response to the natural-language follow-up question; and
wherein submitting the natural-language follow-up question to the generative AI agent comprises resubmitting the two or more initialized context variables to the generative AI agent.
21 . The system of claim 17 , the operations further comprising:
based on the natural language query:
generating a non-LLM query for two or more of the values of metadata associated with the two or more data items of the plurality of data items; and
submitting the non-LLM query to a non-LLM system obtain the two or more of the values of metadata associated with the two or more data items of the plurality of data items; and
training a machine learning model to generate non-LLM queries for values of metadata based on natural language queries.