Every ad you run, graded — your bar, not ours.
Every ad grouped by its creative and graded against your KPI — Winning, Losing, Fatiguing — over 30 days, behind a 10% spend floor. AI vision reads the hook, tone, format, and claims of each.
Live — this is the surface working, on loop
Meta, Google, TikTok, Microsoft — under an hour, no spend access.
Ads grouped by creative; vision reads every asset; grades land against your bar.
Your tags, your vocabulary, your KPI targets — taught once, applied everywhere.
Your creative vocabulary, taught once — read everywhere
Define what “founder tone” means to your team; pick whether each tag reads from the ad name or the creative itself. Watch it read, correct it once — corrections stick.
Applied 3 tags — drafted ad-name rules + read 12 creatives.
Which KPIs can it grade against?+
Any primary KPI you set per brand — ROAS, cost per new customer, CPA. The target is yours; flip it and every grade re-scores.
How does it avoid overcalling small samples?+
A 10% spend floor and explicit low-sample flags. AgentMark reports what has won — never what will.
Does it see creative or just campaigns?+
Creative-level: ads are grouped by their creative asset, so the read is about the idea, not the ad set.
