Exploring Natural Language Features in Model Driven Apps
Natural language processing (NLP) has revolutionized the way we interact with technology, and its integration into model-driven apps is a testament to the advancements in this field. This post explores the history and evolution of natural language features in these apps, highlighting key milestones and the latest innovations that enhance user experience.
A Historical Walkthrough
The journey of natural language features in model-driven apps is marked by continuous innovation and improvement. In 2019, a preview feature allowed users to build Advanced Find Views using natural language — a promising development that, unfortunately, never reached general availability. Despite this setback, the potential of NLP in improving user interaction was undeniable.
Power BI, a tool renowned for its data visualization capabilities, has long incorporated natural language features. The Q&A functionality enables users to ask questions of the data and generate charts based on their queries. This early adoption of NLP in Power BI set the stage for more sophisticated integrations in other applications, including model-driven apps.
Fast forward to the present, and we see the impact of large language models transforming the landscape. Natural language features are now ubiquitous, and Copilot — an AI assistant — is at the forefront of this revolution. The capabilities of Copilot have expanded significantly, offering users unprecedented convenience and efficiency in completing tasks.
Feature: Find Records and Filter Views Using Natural Language
One particularly exciting feature is the ability to find records and filter views using natural language. Leveraging the power of Copilot, users can now find, filter, and sort data more efficiently within model-driven apps by simply describing what they are looking for. This functionality significantly reduces the time previously spent on advanced filtering and enhances overall productivity.
Imagine needing to filter out high-priority cases with overdue follow-up dates. Instead of manually setting up column filters or using advanced filters, you can now describe your request in natural language, and the system will do the rest. This innovation brings a level of convenience and efficiency that was previously unattainable, allowing for more intuitive and user-friendly interaction with data.
With capabilities such as filtering of records, sorting, and text search, this feature is poised to revolutionize the way users engage with model-driven apps. Though currently in preview, it is continually being enhanced with each weekly release, promising even greater functionality in the future. Administrators can enable this feature through the ‘Natural Language Grid and View Search’ setting, opening the door to more advanced and seamless data management experiences.
Smart Grid
Smart Grids are now in preview in model-driven views, allowing users to find, filter, and sort their data quickly with natural language, bypassing complicated advanced filters. If your administrator has turned on Natural Language Grid and View Search, then the natural language search box will be visible.
Natural Language Search
Using smart grid natural language search, you can ask data-related questions with natural language. For instance, requesting “cases with high priority with overdue follow-up by date” will filter your view to display only those relevant cases.
Supported Features
- Filtering of records
- Sorting
- Text search
Known Limitations
The following capabilities aren’t supported:
- Query aggregation
- Query grouping
- Adding columns
Considerations
After executing a query, review the generated filter tags to ensure that the filter conditions were correctly interpreted from your natural language query. If any part of your query is missing from the filter tags, the results weren’t filtered by that condition. If Copilot doesn’t produce the desired results, consider modifying your query by:
- Referring to data columns by their names as they appear in the grid header.
- Separating multiple conditions with commas or periods.
- Search strings with two words or fewer will do a text search (previous functionality). To do a Copilot search, use more than two words. To perform a text search for more than two words, enclose the search term in single or double quotes.
The Future of Natural Language in Model-Driven Apps
The natural language features available in model-driven apps represent a significant leap forward in making technology more accessible and user-friendly. From enhancing productivity with Copilot’s form fill assistance to optimizing query accuracy in the smart grid, these features exemplify the potential of NLP in transforming our digital experiences. As we look to the future, the continued evolution of natural language processing promises to unlock even greater possibilities for model-driven apps and beyond.
