Amazon launched Ads Agent. Here is what it does and does not do.

Amazon's Ads Agent, in beta since Q1 2026, lets you create and adjust campaigns with natural- language prompts instead of manual setup. Early testers report 30–40% time savings and 12–18% ACoS improvements. It is genuinely useful — and it is not the same thing as autonomous, profit-aware bidding. Here is the honest line between what Amazon's native agent does and what it leaves on you.

What is Amazon Ads Agent?

Ads Agent is Amazon's built-in AI campaign assistant. You describe an objective in plain language — "launch a campaign for my new running socks targeting €0.20 ACoS" — and it scaffolds the campaign, suggests keywords, and proposes bid adjustments. It removes the click-work of campaign management. For sellers drowning in manual setup, that is real time back.

What it does well

• Setup and structure. Spinning up campaigns, ad groups, and keyword sets from a prompt is faster than the console.
• Native data access. It sits inside Amazon, so it sees first-party signals without an API handshake.
• Lowering the floor. A novice with Ads Agent will beat a novice without it. That is a meaningful accessibility win.

Where it stops

Three gaps matter, and they are structural rather than temporary:

1. It optimises for the metric you name, not your profit

Tell it to hit an ACoS target and it will — including by pushing bids on keywords that look efficient and quietly erode margin. ACoS is a proxy. Total net profit is the goal, and the two diverge exactly where the money is. We unpack that in ACoS vs TACoS.

2. It is not inventory-aware

Ads Agent will happily bid to first page on an ASIN whose days-of-cover is about to fall off a cliff, because the advertising side does not natively price in your supply side. The stockout, the rank decay, and the expensive recovery that follows are yours. See why your best ASIN is your biggest risk.

3. Its incentive is more spend, not your margin

This is the one nobody says out loud. Amazon's advertising business grows when sellers spend more. A tool built by the marketplace will never be structurally motivated to tell you to spend less. That is not a criticism of the engineering — it is a fact about whose P&L the tool serves.

Ads Agent vs an independent autonomous bidder

The useful framing is not "Amazon's AI vs everyone else's." It is "a campaign assistant vs a profit model." Ads Agent helps you do PPC faster. An independent autonomous system like Mirox is built to do a different job: price every bid against your total net profit, throttle on your inventory, run per-marketplace bid math, and explain every decision in a trace you can audit and hand to finance.

Use Ads Agent to remove busywork. Use an independent profit model to decide how much a click is actually worth to your business. They are not competitors; they answer different questions.

The honest recommendation

If you are a smaller seller doing everything by hand, Ads Agent is a clear upgrade — adopt it. If you are spending five figures a month across marketplaces and your margin is the thing you are protecting, you want a system whose incentive is aligned with yours and whose decisions you can defend. The two can coexist.

See the nine-point checklist for evaluating either, or watch an independent model run in Simulation Mode on your real account before a cent moves.