Naver updated its interactive AI service ‘Clova X’ on the 2nd.
This update introduces a ‘Connector’ feature, allowing users to engage in various conversations based on the content of uploaded document files.
Using the document-based conversation feature in Clova X is straightforward. By uploading a document file in the Clova X conversation input field and entering a question, Clova X provides appropriate answers based on the document’s content. For example, if a user uploads a company’s business report for business analysis, Clova X first summarizes the key points and subsequently offers various responses for drafting a business analysis paper based on the document. This is especially helpful when working with lengthy and complex papers or reports, and can be utilized in daily scenarios, such as conducting mock interviews after uploading a resume file.
Additionally, Clova X has started integration with the mobility platform ‘SoCar’ through a ‘Skill’ system. Skills are features that enhance the user experience by compensating for the limitations of the language model’s generation capabilities in tasks like searching for the latest information, booking places, or purchasing products. With the integration of SoCar’s car-sharing service, after activating the SoCar skill in Clova X, users can input queries related to car rentals, and Clova X will recommend suitable vehicles based on valid SoCar data.
Seong Nak-ho, Head of Naver Cloud’s Hyperscale AI technology, stated, “With Clova X being able to handle lengthy and complex documents, it has evolved into a tool that can provide more assistance to individual productivity.” He added, “We will continue to update the model to derive better outcomes from inquiries and expand the skill ecosystem to enhance the accuracy of responses and convenience of service.”
Meanwhile, since its launch, Clova X has been updating its model performance based on user feedback to improve the quality of responses. Naver reported that the overall quality of responses improved significantly after the update conducted in September, particularly noting enhancements in writing, coding, and summarizing tasks within the productivity domain.
Lee Sang-jin daedusj@autodiary.kr