The problem
What was broken before AI
Documentary teams often work with huge amounts of archival material. For the Muhammad Ali series alone, Florentine Films gathered more than 20,000 still images and hundreds of hours of footage. Every image, clip, audio file, and document needs context: what it shows, where it came from, what metadata it contains, and how it might connect to a story. That logging work is slow, manual, and easy to bottleneck.
What changed
What the use case made possible
Tim turned several of those bottlenecks into custom AI tools. A backend workflow uses vision models, metadata, transcription, and embeddings to catalog archival media. A custom iOS app called Flip Flop helps researchers photograph the front and back of physical archive materials while embedding descriptions and transcripts into the files. A Mac utility called OCR Party lets researchers select a specific region of a difficult document and extract cleaner text from it.
Why this matters
Why this use case is worth studying
Tim’s workflow is valuable because it uses AI to amplify research, not replace it. The creative work is still finding the story, making choices, and building the film. AI helps with the repetitive work that gets in the way: describing images, pairing front/back notes, transcribing audio, making documents searchable, and finding similar assets across a huge archive. That is a practical model for any field built on mountains of source material.
Use this when
When this pattern applies
Use this pattern when a creative or research team has too much source material to describe, tag, and search manually. It works especially well for archives, documentaries, museums, legal collections, historical research, oral histories, and any workflow where real material needs to become searchable without losing context.
