Traditional methodology development relies on single-author cycles spanning years, producing static documents that readers consume passively. This paper presents a design science investigation into AI-Native eBook production — a paradigm in which methodology artifacts are co-created through orchestrated multi-IDE environments (Claude Code, Cursor, Antigravity, Codex), version-controlled in Git, published with persistent identifiers (DOI), and consumed through reader-IDE interaction rather than linear reading. We instantiate this paradigm in the GenAI Consulting Methodology Toolkit (Apache 2.0; n = 354 documents; 7 languages; 22 specialized AI-IDE workflows across three IDE families). In our observed production of this artifact, we find four properties not achieved in prior single-tool authorship attempts within the same project: (1) simultaneous multilingual coherence enforced cross-language, (2) multi-engine adversarial review, (3) reader-queryable execution — the same IDE that produced the artifact can interrogate it through workflows such as /socratic-challenge and /deep-synthesize, and (4) cryptographically reproducible derivation through Git history plus declarative workflow files. We evaluate against Hevner's seven design science guidelines and contrast quantitatively with traditional methodology development cycles (McKinsey 7-Step, Rosemann BPM development). We discuss implications for methodology engineering, the HCI thesis of IDE-as-medium, and the future of active academic publications in the LLM era. This deposit accompanies the GenAI Consulting Methodology Toolkit software deposit (DOI: 10.5281/zenodo.20261680) — the artifact described and evaluated in this paper. See "Related identifiers" below.
Yeh-Hsing Lu (Sun,) studied this question.