How it works

The daily Runbook, end to end.

One scheduled cycle a day. You stay in control at the Approval Gate; everything before and after is automated, bounded by your Guardrails, and fully auditable.

1

Pull

Every morning AdsPlane pulls your latest Amazon Ads reports — Sponsored Products, Brands, and Display — for the campaigns on your allowlist. Your Amazon credentials stay encrypted, and raw reports are processed for the run, not hoarded beyond your retention window.

2

Analyze

The deterministic engine classifies each campaign by role and budget lane, scores keywords and search terms, factors in your product economics and — on higher tiers — AMC path-to-conversion, then decides what (if anything) should change. Python logic makes every decision; AI only narrates it in plain English.

3

Manifest

Recommendations are assembled into a Manifest — a structured, reviewable action plan. Your Guardrail Policy bounds it: max bid and budget moves, protected terms, campaigns and ASINs, confidence thresholds, and daily spend caps. Anything outside the rails is blocked or flagged, never silently applied.

4

Approve

Nothing touches your account until you approve. Review the Manifest in the dashboard, or approve and reject straight from Telegram — everything, a high-confidence subset, or individual actions. This is the Approval Gate, and it is your control point.

5

Execute

Approved actions apply to Amazon through the official Ads API, with a before/after snapshot captured for every change and idempotent, retry-safe execution. Errors fail closed — a run pauses rather than guessing.

6

Reconcile

After execution AdsPlane verifies that each change landed as intended, flags any drift, and writes everything to the Execution Ledger. You get a daily digest with a QC report attached, so you always know what happened and why.

Your control points

Where you stay in control.

Guardrail Policy

Versioned, seller-owned limits on bids, budgets, pauses, negatives, protected items, and confidence — the engine can never act outside them.

Approval Gate

No execution without your approval — via the dashboard or Telegram. Start in Shadow Mode to preview Manifests with no execution at all.

Emergency stop

Pause everything for your account in one click, any time. Automation fails closed: on error it stops, it never continues blindly.

When you’re ready

Autonomy is earned, never assumed.

Every seller starts at Shadow Mode or Manual Gate. As clean Reconciliation runs accumulate and you build trust, you can opt into guided auto-approval for low-risk actions — always within your Guardrails, and never an AI deciding to spend. See the Autonomy Ladder →

Ready to see it on your account? Start a free trial → or compare plans →