Who Asked for This? Inline Annotations as Authoring Transactions for Provenance in Agentic Authoring

2026-09-30

AI Summary

This study addresses the challenge of tracing correspondences between user requests and document modifications in AI-assisted writing by proposing the Reactant paradigm. This approach introduces a novel inline annotation interaction mechanism grounded in a validated transaction protocol. By integrating a state witness kernel with word-level alignment techniques, it transforms opaque generative processes into verifiable structured records, establishing a token-level vertical lineage that enables precise tracking of insertions, deletions, and transformations. The system has been successfully deployed in academic paper revision and daily collaborative writing scenarios, supporting historical retrospective querying and skill reuse extraction. These applications effectively validate its practicality and scalability within agent-based collaborative environments.

Abstract

Writing with AI agents turns a paragraph into the outcome of many requests, yet the finished document rarely explains which request produced which change. We introduce Reactant, an interaction paradigm in which authors place typed inline annotations in their original documents. A verified transaction protocol records each request, skill identity, and the correspondences that identify additions, deletions, and transformations. The kernel validates the witness against the recorded states to establish word-level longitudinal lineage. We demonstrate Reactant through this paper's revision history, and report four months of three colleagues' self-directed use. Their uses include conversational inquiry into history and deriving reusable skills from recurring requests, illustrating how the transaction record serves as an extensible substrate for agentic authoring.

Problem

• Agentic Authoring
• Provenance
• AI-assisted Writing
• Lineage Tracking

Innovation

• Agentic Authoring
• Inline Annotations
• Transaction Protocol
• Provenance
• Longitudinal Lineage