TaskArtisan: Designing Composable Generative Widgets for LLM-Assisted Analysis

2026-07-19

AI Summary

This work addresses the limitations of unstructured, linear interactions in large language model (LLM)-assisted analysis, which hinder workflow maintainability and efficiency due to the absence of navigational and backtracking mechanisms. We introduce TaskArtisan, a technical probe enabling users to create and compose generative analytical UI components, representing the first systematic exploration of generative user interfaces in analytical tasks. Through LLM-generated high-fidelity GUI code, user interviews, tool analysis, and a controlled study (N=12), we formulate a composable generative widget paradigm and derive a design framework centered on malleability, prescriptiveness, and interoperability. Our findings demonstrate that this approach significantly enhances the clarity and visual expressiveness of analytical workflows, while also revealing inherent tensions—such as rigidity and prompt complexity—introduced by generative UIs, thereby establishing a foundational trade-off framework to guide future design.

Abstract

People increasingly use chatbots such as ChatGPT for everyday analysis tasks. While chatbots unify many analysis functions (e.g., scripts, visualizations, summaries), long conversations become hard to navigate, making it difficult to revisit prior steps or reuse successful workflows. LLMs now generate high-fidelity GUI code that enables people to create customized analysis tools beyond text. Yet, what new opportunities generative UIs bring to analysis work remain unclear. We interviewed six professionals about analysis with chatbots, analyzed publicly shared LLM-generated GUI tools, and conducted a comparison study (N=12) between a chatbot and TaskArtisan, a technology probe that enables people to create and assemble generative analysis UI widgets for sequential and fan-out composition. We find that GUI improved clarity and visual presentation but also introduced rigidity and additional prompting challenges. We summarize the trade-offs into a provisional design framework (malleability, specification, interoperability) to inform future generative UI in LLM-assisted analysis workflows.

Problem

• LLM-assisted analysis
• generative UI
• analysis workflows
• composable widgets
• chatbot interaction

Innovation

• generative UI
• composable widgets
• LLM-assisted analysis
• TaskArtisan
• design framework