TableTalk: Scaffolding Spreadsheet Development with a Language Agent

TL;DR

TableTalk is a language agent that scaffolds, guides, and incrementally builds spreadsheets, improving quality and reducing cognitive load for programmers by leveraging professional workflows and AI tools.

Contribution

We introduce TableTalk, a novel agentic system that enhances spreadsheet programming through structured scaffolding, flexible planning, and incremental construction based on empirical studies.

Findings

TableTalk increased spreadsheet quality and user preference by 2.3 times.

It reduced cognitive load and thinking time by 12.6%.

The system offers structured guidance aligned with professional workflows.

Abstract

Spreadsheet programming is challenging. Programmers use spreadsheet programming knowledge (e.g., formulas) and problem-solving skills to combine actions into complex tasks. Advancements in large language models have introduced language agents that observe, plan, and perform tasks, showing promise for spreadsheet creation. We present TableTalk, a spreadsheet programming agent embodying three design principles -- scaffolding, flexibility, and incrementality -- derived from studies with seven spreadsheet programmers and 85 Excel templates. TableTalk guides programmers through structured plans based on professional workflows, generating three potential next steps to adapt plans to programmer needs. It uses pre-defined tools to generate spreadsheet components and incrementally build spreadsheets. In a study with 20 programmers, TableTalk produced higher-quality spreadsheets 2.3 times more likely to be preferred than the baseline. It reduced cognitive load and thinking time by 12.6%. From this, we derive design guidelines for agentic spreadsheet programming tools and discuss implications on spreadsheet programming, end-user programming, AI-assisted programming, and human-agent collaboration.