Data-driven articles combine narrative text with visualizations to communicate insights but are time-consuming to produce, requiring both narrative composition and accurate text–chart linking. Existing automated linking methods struggle with accuracy, and automated narrative generation often misaligns with user intent and may contain hallucinations. We present VisCopilot, which leverages retrieval-augmented generation and large language models to provide fact-based, intent-aligned suggestions. As users write, the system supports narrative completion and text–chart linking in a single step, with a human-in-the-loop approach ensuring reliability. Technical evaluation on a benchmark dataset of data-driven articles authored by journalists confirmed the accuracy and quality of VisCopilot’s links. A user study further showed that VisCopilot outperforms baselines in usability, accuracy, and user confidence, with participants valuing its fact-based autocompletes for aligning suggestions with their writing intent. VisCopilot contributes: (1) an autocomplete-based design for assisting data-driven article authoring that balances efficiency, accuracy, and user autonomy; and (2) a reframing of text–chart link generation as a data-fact recommendation problem, enabling intent-aligned retrieval and systematic evaluation.

Recent advances in open-domain question answering over tables have widely adopted large language models (LLMs) under the Retriever-Reader architecture. Prior works have effectively leveraged LLMs to tackle the complex reasoning demands of the Reader component, such as text-to-text, text-to-SQL, and multi-hop reasoning. In contrast, the Retriever component has primarily focused on optimizing the query representation—training retrievers to retrieve relevant tables based on questions, or to select keywords from questions for matching table segments. However, little attention has been given to enhancing how tables themselves are represented in embedding space to better align with questions. To address this, we propose QGpT (Question Generation from Partial Tables), a simple yet effective method that uses an LLM to generate synthetic questions based on small portions of a table. These questions are generated to simulate how a user might query the content of the table currently under consideration. The generated questions are then jointly embedded with the partial table segments used for generation, enhancing semantic alignment with user queries. Without the need to embed entire tables, our method significantly improves retrieval performance across multiple benchmarks for both dense and late-interaction retrievers.