Jean-François Fèvre + Gilles SABART
The Jean-François Fèvre and Gilles SABART collaboration is a 2024-2025 working relationship in which Steven attempted to turn AI transcription, interview analysis, legal-report generation, and document synthesis into a usable professional workflow for internal investigations and related legal work. The associated project name is Ainternals.
Structure
Jean-François Fèvre appears as the primary operational requester and field investigator. He sends audio recordings, tracks dossiers, gives feedback on transcripts and summaries, and pushes for useful deliverables.
Gilles SABART appears as the lawyer and process/legal authority. He is involved in report templates, legal framing, compliance questions, training workflows, and later direct use of Steven’s tools and advice.
Steven is the technical operator and proto-product builder: he receives recordings, produces transcripts, summaries, timelines, analyses, shared folders, Notion pages, and prototypes for transcript editing and workflow automation.
Main workflow
The collaboration covers several work surfaces:
- interview transcription from audio recordings
- correction and cleanup of transcripts
- summaries of interviews
- legal/investigation report templates
- timelines across transcripts and annexes
- analysis of documents and annexes
- shared Drive/Notion organization
- early transcript-editor and upload tooling
- AI-assisted legal/investigation methodology
The group thread describes the workflow as a mix between Steven running the “machine” and Jean-François/Gilles doing human synthesis, legal correction, and final professional judgment.
Dossiers
The messages mention several dossiers or working labels, including Wessling, Lidal, Clufix, Gosset, SOCARA, and Innergex. These serve as real-world test cases for the tooling, not abstract demos.
Product pressure
The relationship gradually shifts from manual service to software expectation. By late 2024 and early 2025, Jean-François and Gilles are no longer only asking for transcripts and summaries. They are waiting for an application or interface that would make the workflow repeatable and usable without Steven manually doing every step.
The project, Ainternals, therefore becomes an early proof that Steven could identify and build around a real professional pain. It also exposes the gap between prototype value and operational reliability.
Breakdown
In early 2025, the collaboration strains under delays, missed contact, financial pressure, and unfinished product delivery. Jean-François asks whether they can still count on Steven. Gilles separately states that the application must be delivered cleanly before broader support can be discussed.
The rupture is not a simple failure of usefulness. The messages show that the outputs had value, the product direction was plausible, and the clients wanted it to work. The failure point was continuity: delivery, responsiveness, payment expectations, and Steven’s shutdown/executive dysfunction under load.
Significance
This collaboration is one of the clearest bridges between Steven’s private AI cognitive-scaffold work and a real external service/product attempt. It belongs in the wiki as both a relationship cluster and a project-like episode.