The Challenge
Nonfiction TV producers and documentary filmmakers face a structurally slow development process — pitching a new season requires weeks of research, story sourcing, and pre-production prep before a network will greenlight anything. For production companies in a contracting content market where budgets have shrunk and buyers are harder to move, this lag kills momentum and puts smaller teams at a severe disadvantage against better-resourced studios.
What They Built
An Australian sports documentary producer fed all Season 1 episode transcripts into Claude with structured prompts, generating a complete Season 2 treatment that synthesized existing narrative patterns, characters, and story logic — delivering a polished pitch document to the network the same day.
Fred Grinstein began with a clear constraint: the producer needed to pitch a Season 2 treatment but couldn't afford weeks of research in a contracting content market. The existing asset was Season 1 — a full library of episode transcripts containing narrative logic, character arcs, and story patterns that hadn't been systematically mined. The approach was structured prompting: all Season 1 transcripts were loaded into Claude, with prompts designed to surface repeating story structures, key characters, and thematic threads that would logically extend into a second season. The AI synthesized across the full transcript corpus and generated a compelling Season 2 treatment in hours. The producer reviewed, refined, and delivered the pitch document to the network the same day. The network responded positively and moved toward proper development. The broader implication was methodological: this wasn't a one-off shortcut but a validated framework — any nonfiction producer with an existing transcript library can apply the same approach to accelerate pitch development for sequels or spin-offs.
AI Role
AI-generated Season 2 treatment was strong enough for the network to respond positively and advance to proper development — proving AI-assisted pitching can satisfy buyer expectations at the pre-greenlight stage.
Infrastructure
• Claude (Anthropic) — transcript synthesis and treatment generation • ChatGPT — supplementary content development • Google Gemini — supporting research and drafting • Midjourney — visual development • Perplexity — research augmentation • CapCut — video/edit support
Integration Points
• Season 1 episode transcripts → Claude with structured prompts → narrative pattern extraction • Extracted patterns → Claude synthesis → Season 2 treatment draft • Treatment draft → producer review and refinement → network pitch submission • Supplementary tools (ChatGPT, Gemini, Perplexity) → supporting research and content development
