The why-did-it-move investigation engine.
Ask any metric question; the agent decomposes drivers, tests hypotheses across every connected source, and ranks the root causes by dollar impact.
Four ways it earns its place in your stack.
-
Driver decomposition
Every metric move broken into ad spend, returns, COGS and pricing buckets, sized in dollars.
-
Hypothesis testing
Drivers labeled CONFIRMED, TESTING or REFUTED with cited evidence from every connected source.
-
Dollar-ranked causes
Root causes sorted by their actual contribution-margin impact, never alphabetical or arbitrary.
-
Reproducible runs
Every investigation has a permalink, a credit cost, and a re-run button if a source resyncs later.
A live agent timeline, an evidence ledger, a sized answer.
- Run timeline walks through pulling Shopify orders, fanning out queries, reconciling fees and returns, and writing the report.
- Driver buckets, Returns, Ad spend, COGS, Pricing, each tagged with confidence and citation count.
- Final answer hands off straight to a recommended fix in the optimization queue.
One brain across your whole business.
Pillars are one connected loop, explore the rest of the platform.
See your hidden profit in one session.
Connect your stack, run your first investigation, and watch the agent find money your dashboards never could.
~14 day payback · no card to start · pause anytime.