Composable Prompting Workspaces for Creative Writing - Exploration and Iteration Using Dynamic Widgets

Sep 15, 2026

Uploaded by Skyler

Generative AI models enable flexible text creation and transformation, yet standard graphical user interfaces for prompting restrict iterative exploration because prompts remain static text fields. The work introduces a composable prompting canvas that uses dynamic, persistent widgets to capture task-relevant facets. Widgets can be generated via system suggestions, direct prompting, or manual creation, supporting user-driven customization. A user study with 18 participants shows higher control, better Creativity Support Index results than a conversational baseline, and strong preference for the proposed interface.

LanguageEnglish

CategoryPresentations

FormatPDF

Pages11

Reading Time4 min

AccessFree to download

Introduction

Motivation and limitations of current GUIs

PromptCanvas and dynamic widgets

arXiv :2503 .21394v1 [ cs .HC] 27 Mar 2025 Composable Prompting Workspaces for Creative Writing: Exploration and Iteration Using Dynamic Widgets Rifat Mehreen Amin LMU Munich Munich, Germany [rifat.amin@ifi.lmu.de](rifat.amin@ifi.lmu.de) Oliver Hans Kühle LMU Munich Munich, Germany [o.kuehle@campus.lmu.de](o.kuehle@campus.lmu.de) Daniel Buschek University of Bayreuth Bayreuth, Germany daniel.buschek@uni[bayreuth.de](bayreuth.de) Andreas Butz LMU Munich Munich, Germany [andreas.butz@ifi.lmu.de](andreas.butz@ifi.lmu.de) Figure 1: User interface of PromptCanvas. (A) Canvas-like workspace where users can place and organize widgets to create a customized environment. (B) Widget panel. (b1) Example of widgets created through system suggestions. (b2) Button to get widget suggestions from the system. (b3) Field for entering prompts to create multiple widgets of a specific theme. (b4) Example of an empty widget created by double-clicking at empty space. (C) Text editor and output text generations from the prompt.(c1) Button to rephrase the text based on the widgets on the canvas (light blue). (c2) Field to provide prompts for generating text. (D) Example of an opened widget with suggested values for customization. In each widget, users can request more such suggestions and save current input for refinement and iteration. (E) Menu bar for creating, copying, or deleting a canvas. Abstract Generative AI models offer many possibilities for text creation and transformation. Current graphical user interfaces (GUIs) for prompting them lack support for iterative exploration, as they do not represent prompts as actionable interface objects. We propose the concept of a composable prompting canvas for text exploration and iteration using dynamic widgets. Users generate widgets through system suggestions, prompting, or manually to capture task-relevant facets that affect the generated text. In a comparative Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s) . CHI EA’25, Yokohama, Japan © 2025 Copyright held by the owner/author(s) . ACM ISBN 979-8-4007-1395-8/2025/04 [https://doi.org/10.1145/3706599.3720243](https://doi.org/10.1145/3706599.3720243) study with a baseline (conversational UI), 18 participants worked on two writing tasks, creating diverse prompting environments with custom widgets and spatial layouts. They reported having more control over the generated text and preferred our system over the baseline. Our design significantly outperformed the baseline on the Creativity Support Index, and participants felt the results were worth the effort. This work highlights the need for GUIs that support user-driven customization and (re-)structuring to increase both the flexibility and efficiency of prompting. CCS Concepts • Human-centered computing → Interactive systems and tools; Graphical user interfaces. Keywords Dynamic UI, Prompting, LLM, human-AI co-creation, creativity support ACM Reference Format: Rifat Mehreen Amin, Oliver Hans Kühle, Daniel Buschek, and Andreas Butz. 2025. Composable Prompting Workspaces for Creative Writing: Exploration and Iteration Using Dynamic Widgets. In Extended Abstracts of the CHI Conference on Human Factors in Computing Systems (CHI EA’25), April 26-May 1, 2025, Yokohama, Japan. ACM, New York, NY, USA, 11 pages. [https://doi.org/10.1145/3706599.3720243](https://doi.org/10.1145/3706599.3720243) 1 Introduction Advancements in generative AI models have revolutionized text interaction, offering powerful tools for creating and exploring text [23, 24, 28] . These may enhance creative expression by providing users with nove

FAQs

What limitation of current prompting GUIs does the document address?

Current GUIs present prompts as static text fields, which limits iterative exploration and customization needed for creative writing workflows.

How does the proposed PromptCanvas help users with prompting?

PromptCanvas turns prompts into actionable, persistent interface objects by letting users create and arrange dynamic widgets on a canvas for iterative refinement.

What were the main findings from the user study?

In a study with 18 participants, users reported more control and preferred PromptCanvas over a baseline, and the system significantly outperformed it on the Creativity Support Index.

Sep 15, 2026

Uploaded by Skyler