Abstract: Implementations relate to leveraging large language model(s) (LLMs) and vision language model(s) (VLMs) to facilitate human-to-computer dialogs. In various implementations, one or more digital images may be processed using one or more VLMs to generate VLM output indicative of a state of an environment. An LLM prompt may be assembled based on the VLM output and a natural language input. The LLM prompt may be processed using one or more LLMs to generate content that is responsive to the natural language input. The content that is responsive to the natural language input may subsequently be rendered at one or more output devices.

Abstract: Searching and locating an event on a calendar with a timeline is provided. An application such as a calendar application detects a search on the calendar or the timeline. One or more events are located on the calendar or on the timeline that match a search term of the search. The search term on the one or more events is highlighted. The one or more events with the search term that is highlighted are displayed on the calendar or the timeline.

Abstract: Techniques enable contextually-aware selections of event forums. Scheduling data can be analyzed to identify a context associated with an appointment. A forum that is appropriate for the identified context can be selected and utilized. The analysis can also utilize traffic data, map data, weather data, user preferences, and other contextual data to select one or more forums. A forum can be a place or communication medium. A service provider can associate one or more forums with a particular skill set. For example, a legal office can associate an online forum, such a videoconference, with a first category of services, and an on-site location for a second category of services. Depending on a category of services identified in one or more user preferences and/or meeting requests, a forum can be selected. Other factors, which may include weather conditions, traffic conditions, and commute times can also influence the selection of a forum.

Abstract: Locating an event on a timeline is provided. A calendar application may locate an event on a timeline by detecting a selection action on the event on a calendar. A context menu is displayed on the calendar, where the context menu includes a search control. An activation of the search control is detected. The event is located on the timeline of the calendar. The event is displayed on the timeline.

Abstract: A for-your-information event can be received in a computer system. The for-your-information event can include an event time field that indicates a scheduled time for the for-your-information event. The computer system may not automatically block the scheduled time of the for-your-information event on a calendar in response to receiving the for-your-information event. User input indicating that the for-your-information event is to be converted to a regular event can be received. In response to the user input, the computer system can convert the for-your-information event to a regular event. Converting the for-your information event can include automatically including the scheduled time of the for-your-information event as a scheduled time for the converted regular event. Also in response to the user input, the computer system can automatically block the scheduled time of the for-your-information event on the calendar.

Abstract: Techniques described herein provide resource-based auto completion for data entry fields of a graphical user interface. In some configurations, the techniques disclosed herein utilize input data and contextual data from a number of resources to select and display text entry candidates for assisting users in populating one or more data entry fields. For example, the input data can include an indication of a service category or a topic and other data defining a calendar event. The contextual data can be related to service providers and customers including, data defining a prior work history between two or more entities, commute projections, scheduling conflicts, preferences, and other data, such as traffic data and weather data. Data indicating a lifetime value of a customer or provider can also be utilized to select text entry candidates.

Abstract: Techniques described herein provide mechanisms for generating contextually-aware scheduling exceptions. In some configurations, when a scheduling conflict is detected, the techniques disclosed herein can utilize contextual data from a number of resources to determine if a scheduling exception can be made. The contextual data can include preferences, such as preferences of a service provider or a customer, that define criteria and/or goals. The techniques disclosed herein prioritize customers based on the contextual data and provide different scheduling options for customers and other entities based on a priority associated with individual customers. When there is a conflict between two or more calendar events, a scheduling exception can be made for some customers and a scheduling conflict can be made for other customers depending on one or more priorities associated with the customers.

Abstract: Techniques enable contextually-aware selections of event forums. Scheduling data can be analyzed to identify a context associated with an appointment. A forum that is appropriate for the identified context can be selected and utilized. The analysis can also utilize traffic data, map data, weather data, user preferences, and other contextual data to select one or more forums. A forum can be a place or communication medium. A service provider can associate one or more forums with a particular skill set. For example, a legal office can associate an online forum, such a videoconference, with a first category of services, and an on-site location for a second category of services. Depending on a category of services identified in one or more user preferences and/or meeting requests, a forum can be selected. Other factors, which may include weather conditions, traffic conditions, and commute times can also influence the selection of a forum.

Abstract: Techniques provide a contextually-aware resource manager. In response to one or more events, such as the creation or modification of a calendar event, one or more contextually-aware recommendations are generated and displayed to a user. For example, a recommendation can include the names of service providers, the names of customers, time slots for one or more calendar events, and notifications of one or more conditions. The recommendation can be based on data defining a level of eligibility for service providers and customers. The level of eligibility can be determined by a wide range of contextual data, including but not limited to traffic data, payment data, location data, map data, preference data, scheduling data, workload data, work history data, status data, skill set data, or weather data. The techniques assist user interaction with a computing device, and among other benefits, saves computing resources and reduce the number of inadvertent user entries.

Abstract: Techniques described herein provide contextually-aware insights into calendar events. Generally described, the techniques disclosed herein can analyze a wide variety of contextual data including, but not limited to, weather data, traffic data, location data, performance data, preference data, and scheduling data, to generate salient insights that can be automatically displayed and/or communicated to a user. Insights related to one or more calendar events may be generated in response to a discovery of a predetermined condition. A predetermined condition may be detected at the time an appointment is made or at a later time when contextual data indicates a change in one or more conditions. An insight can include a text description, an image, a graphical indicator, a generated voice, and any other suitable form of communication describing useful information regarding one or more calendar events. An insight can include ranked list of recommendations can also be displayed.

Abstract: Techniques described herein provide the automation of workflow events. Input data can be processed with contextual data to generate workflow data. Workflow data can define a multi-stage process describing aspects of an orchestrated pattern of activity. In some configurations, input data indicating a calendar event can be received by one or more computing devices. The input data can include a voice input or text description indicating a desire to initiate a project, define a goal, achieve one or more metrics, etc. In response to receiving the input data, contextual data can be received from a number of resources, including specialized resources having a knowledgebase for particular topics. The techniques disclosed herein utilize the contextual data to generate workflow data that can define a multi-stage process. Calendar events, reminders and other forms of communication can be generated from the workflow data.

Abstract: Sharing a timeline of a calendar is provided. An application such as a calendar application detects an activation of a share operations control on a timeline of a calendar. A context menu is displayed on the timeline, where the context menu includes a share control and a print control. An activation of the share control is detected. A communication component is initiated to share the timeline.

Abstract: A for-your-information event can be received in a computer system. The for-your-information event can include an event time field that indicates a scheduled time for the for-your-information event. The computer system may not automatically block the scheduled time of the for-your-information event on a calendar in response to receiving the for-your-information event. User input indicating that the for-your-information event is to be converted to a regular event can be received. In response to the user input, the computer system can convert the for-your-information event to a regular event. Converting the for-your information event can include automatically including the scheduled time of the for-your-information event as a scheduled time for the converted regular event. Also in response to the user input, the computer system can automatically block the scheduled time of the for-your-information event on the calendar.

Abstract: Visualizing events on a calendar with a timeline is provided. An application such as a calendar application receives a request to associate a calendar with a timeline. Events are detected on the calendar to present on the timeline. The events are displayed on the timeline, where the timeline is placed in a location that is below the calendar.

Abstract: Sharing a timeline of a calendar is provided. An application such as a calendar application detects an activation of a share operations control on a timeline of a calendar. A context menu is displayed on the timeline, where the context menu includes a share control and a print control. An activation of the share control is detected. A communication component is initiated to share the timeline.

Abstract: Creation of a calendar event from a timeline is provided. An application such as a calendar application detects a create action to create an event on the timeline. Details of the event are captured that includes a subject name of the event and an event date. The event is displayed on the timeline based on the details. A save action is detected to save the event. The event is displayed on a calendar associated with the timeline based on the details.

Abstract: Locating an event on a timeline is provided. A calendar application may locate an event on a timeline by detecting a selection action on the event on a calendar. A context menu is displayed on the calendar, where the context menu includes a search control. An activation of the search control is detected. The event is located on the timeline of the calendar. The event is displayed on the timeline.

Abstract: Searching and locating an event on a calendar with a timeline is provided. An application such as a calendar application detects a search on the calendar or the timeline. One or more events are located on the calendar or on the timeline that match a search term of the search. The search term on the one or more events is highlighted. The one or more events with the search term that is highlighted are displayed on the calendar or the timeline.

Abstract: Management of an event with a content on a calendar with a timeline is provided. An application such as a calendar application detects a selection of the event with the content on the calendar. A context menu is displayed on the calendar, where the context menu includes an add control. An activation of the add control is detected. The event with the content is duplicated on the timeline of the calendar.

Abstract: Management of an event on a calendar with a timeline is provided. An application such as a calendar application displays an initial context menu that includes a create control on the calendar, in response to an initial selection action of the event displayed on the calendar. The event is presented on the timeline, in response to an activation of the create control. A new context menu, that includes a delete control, is displayed on the timeline, in response to a new selection action of the event displayed on the timeline. The event is removed from the timeline, in response to an activation of the delete control.