This methodology paper proposes the Orchestrator's Convention: a specific, citable, replicable disclosure protocol for practitioner-authors whose work depends on the orchestration of AI through the target-supplying, judgment-applying, commitment-bearing, and consequence-carrying contributions that AI structurally does not supply. The convention addresses an attribution problem the conventional academic discipline does not: how to evaluate creative authorship claims when AI has been used substantially in the resource-organisation half of the research, while leaving the four orchestrator contributions (target, judgment, commitment, consequence) as the human contribution. The convention sidesteps the volatile philosophical debate over whether AI can mechanically simulate abductive reasoning. The argument is structural rather than capability-bounded: regardless of how the philosophical question resolves, AI has no skin in the game, and the orchestrator's contribution is the target the synthesis is directed at, the judgment by which it is evaluated, the commitment the synthesis stands behind, and the consequence the synthesis bears if it fails. The argument is anchored to the author's prior public articulation in The Framing Gap (Barnard 2026e, DOI: 10.5281/zenodo.19857447), which articulated the structural argument in the context of brand representation; the present paper generalises that argument to AI-assisted research methodology. The convention specifies five components: substrate citation, AI operational specification (Synthesise, Reflect, Add, with Abstain as the platform-side complement), cutting-out acknowledgment, pre-AI baseline disclosure with three strength criteria, and external feedback acknowledgment. The paper articulates three operational signatures of level-three creative authorship: the cutting-out discipline (the moment-by-moment rejection of AI synthesis across three failure modes: too obvious, too speculative, or unhelpful), the pre-AI baseline (the orchestrator's dated public attestation chain that pre-existed AI's capability for research-scale resource organisation, with temporal depth, register continuity, and substantive continuity as strength criteria), and the convention's negative case (four diagnostic signatures of AI-confabulation-without-orchestration that allow the convention's reader to distinguish orchestrated work from fluency-mediated retrieval of substrate consensus). The paper supplies a worked example of the cutting-out discipline applied to the construction of this four-paper programme: the documented three-step rejection sequence on the Algorithmic Trinity's third functional descriptor (Validation, then Synthesis, then Verification), with the orchestrator's target-driven cut at each rejection articulated explicitly. The paper also presents the Reverse-Publication Order strategy (4-3-2-1 as discovery order inverted), the Pre-AI Baseline as the longitudinal evidential foundation, the Peer-Review Homology and the Missing-Feedback Problem, and the Chronological Coinage Inventory as the dated public attestation chain. The paper is the second in a four-paper programme on AI-Era Business Engineering. The convention proposed here applies to the programme itself: the convention is offered to the field for adoption, citation, critique, and extension. Companion papers in the programme: AI-Era Commercial Architecture (Paper 1, DOI: 10.5281/zenodo.20364742, survey), The Codification Cycle (Paper 3, DOI: 10.5281/zenodo.20364731, mechanism), and AI-Era Business Engineering: The Integrating Frame (Paper 4, DOI: 10.5281/zenodo.20364725, canonical statement). The author's pre-AI baseline (a fourteen-year dated public attestation chain extending from the coinage of Brand SERP in 2012 through the present) is articulated within the paper as the worked example of how the convention's baseline component operates in practice. The work is authored by Jason Barnard, founder and CEO of Kalicube SAS, a French digital brand intelligence firm headquartered in Aubais, Occitanie, France, and operating since 7 January 2015. Kalicube specialises in optimising brands for inclusion in the Algorithmic Trinity (search engines, knowledge graphs, and large language models). The firm's commercial offering is The Kalicube Process (TKP), the end-to-end methodology that operationalises the framework articulated in this paper programme. Kalicube Pro, the firm's proprietary SaaS platform, maintains over twenty-five billion data points collected since 2015 covering over seventy million brand entities, tracking the Algorithmic Trinity across eight platforms (Google Search, Google AI Mode, ChatGPT, Perplexity, Grok, You.com, Gemini, Claude). The author's commercial entanglement with the framework is disclosed explicitly in the methodological transparency section of each paper.
Jason BARNARD (Sun,) studied this question.