How does AI video creation
from a brief work?
You write the brief. Picco-AI shoots the film, in three phases like a real production: plan, generate, cut and export. This page explains what happens in each phase, how long it takes and which models do the work.
AI video creation from a brief means: you describe the project in plain language, with topic, goal, audience, mood and length. A pipeline performs the production steps a team would otherwise handle. You write the brief, Picco-AI shoots the film. Phase 1 plans in four stages: analysis of the brief, concept directions, concept bible, shot list. Phase 2 generates a still, a clip, voice-over and music for every shot. Phase 3 edits into a finished MP4 or exports an edit package for Final Cut Pro, Premiere Pro and DaVinci Resolve. Unlike a prompt tool, you write no prompt at all. A quality scout checks every stage before the next one starts, and you choose whether the pipeline runs through or waits for your sign-off after each stage.
What happens in Phase 1 — planning?
Phase 1 is the production office. Four stages build on each other, and a quality scout checks each one. In marketing projects stage 1 is called brief catch, in film and animation project catch; the logic is the same.
Analysis with extended thinking: goal, audience, tone, mandatory content, length. Gaps are named rather than guessed.
Several concepts, each with a tension arc and a performance hypothesis. You pick one.
Characters, look, palette, sound, voice: the binding reference for every shot. For animation with character sheets and style lock.
Every shot with image description, camera, duration, model routing and voice-over text. A validator checks rules and tension arc and repairs automatically.
What happens in Phase 2 — generation?
Phase 2 produces the media, shot by shot: the still via Nano Banana 2, from it the clip via Veo 3.1, plus voice-over and music. Everything runs in one pass; if something is missing, the same run fetches it. With uploaded photos of people, every output is checked against the reference; whatever does not come close to the identity is tried again. You assign your own footage to shots here; from then on it behaves like an AI clip.
What happens in Phase 3 — edit and export?
Phase 3 turns the media into a result. The direct render delivers a finished MP4 with opening and closing titles and AI disclosure per the EU AI Act and, on request, an upscaled 4K master without a new generation run. The edit package comes with it: a timeline for Final Cut Pro, Premiere Pro or DaVinci Resolve with all media on their own tracks, checked by the import simulator.
How long does a project take from brief to video?
From internal test runs in September 2026 with 14 to 16 shots: Phase 1 about 45 to 90 minutes, Phase 2 about 20 to 40 minutes, Phase 3 a few minutes. The times depend on shot count, clip duration and model tier; in sign-off mode your own response time is added. These are measurements from tests, not a promise.
Which AI models work in the pipeline?
Each task gets its own model. Generation is included in the guide value of the subscription and billed against your quota by usage; you choose resolution and Veo tier per project, and a line shows in advance what that costs.
All statements about what a plan includes are rough guide values without guarantee. The plan description and terms at the time of purchase apply.
Analysis, concepts, bible and shot list run on Claude with extended thinking by default. Optionally a local Ollama model when the brief must not leave the machine.
One still per shot, photorealistic, documentary or in the locked animation style. It is the reference for the clip and the storyboard panel.
Clips of 4 to 8 seconds from still and shot description, 1080p as standard, native 4K optional depending on project settings. Silent refusals are detected and retried with another variant.
Voice-over from the shot text, music matched to mood and length.
How much control do I have over the process?
Three levels, chosen per project. Hands-free: you hand in the brief and collect the video. Sign-off per stage: the pipeline stops after each stage, shows result and rating, you approve or correct. Expert mode: access to every line of the shot list, regeneration of single shots, replacement of voice-over or music.
Questions about the process.
Picco-AI shoots the film.
7 days free — paid plans from €79/month. Start today and see the first result on your own material.
→ Start 7 days freemacOS · Windows · Linux