Microsoft is developing an AI model for Excel spreadsheets, enabling natural language prompts
Microsoft's SpreadsheetLLM AI model enhances Excel with natural language prompts, streamlining data analysis and reducing costs but is still in research phase.
Microsoft has developed a framework called SpreadsheetLLM, leveraging large language models to analyze and interpret spreadsheet data. This model addresses challenges with traditional LLMs in handling extensive grids and varied formatting by serializing data and incorporating cell addresses, values, and formats into a data stream.
The tool comprises three modules: one that analyzes and discards non-table content, another that translates data to a more efficient form, and a third that aggregates data. Despite its current limitations, such as ignoring cell background colors and lacking semantic-based compression, it outperforms traditional approaches by 25.6 percent in GPT-4's in-context learning and reduces token usage by 96 percent with its SheetCompressor.
The model's ability to handle structured and unstructured data may reduce hallucinations in AI-generated outputs. It enables non-technical users to manipulate data with natural language queries, potentially impacting jobs in fields like accounting and finance. However, SpreadsheetLLM is still in the research phase and not ready for integration into Microsoft Excel or other commercial products.