Reading, scoring, and saving research output

One file, scored section by section

The phase produces a single file, and each section inside it carries three things: the data itself, a quality score, and the time it last ran. You can watch the sections fill in while a run is still going, because the viewer renders each one the moment it commits rather than making you wait for the whole phase.

The quality score

Right after a section finishes its main work, Pinnora runs a quick, low-cost automated check and attaches the result. The score is a tier, green, amber, or red, plus a number from 0 to 100 and a short list of flags naming what looks off. A red section is your cue to re-run it, and the re-run folds in the suggested fixes. The section scores roll up into one phase-level summary, so a glance tells you whether the grounding is solid.

Saving is automatic and versioned

You never save research by hand. When a run finishes, the output is pinned as a new version, and the research is also indexed into your organization's brand memory so later creatives can reuse it. Re-running one section writes a new version of that section while carrying the others forward, so you build up a history instead of painting over the old one. It works like autosave that quietly keeps every draft.

Why this matters downstream

This is research once, ground everywhere, in practice. Every image, video, document, and ad reads from this file. Keeping the five sections green is the single most effective thing you can do to make your generated creative specific to your brand instead of generic to everyone's.