Amazon’s latest Seller Assistant announcement should get agency owners’ attention. Amazon is bringing more of a seller’s inventory, pricing, advertising, listings and compliance information into an assistant that can offer recommendations and, with the seller’s permission, take action. Alongside changes to Seller Central, it points towards clients getting answers and completing routine work with less help from an agency.

That does not make agencies redundant. It does make the value of some familiar services harder to explain. If a client can see a problem sooner, generate a report in minutes or act on a recommendation themselves, “we monitor the account” becomes a weaker reason to retain an agency. The stronger question is whether the agency can tell them which action is worth taking, why, and what it means for the business beyond Amazon.

From output to judgement

“For less mature agencies, the shift is in where the value sits. As AI takes on more routine work, the question becomes whether the agency understands why it made a recommendation and what that decision means for the client’s margin and longer-term strategy. There’s an opportunity to spend less time assembling reports and more time looking across the account, spotting patterns and building a strategic view. But agencies that haven’t developed that capability may find themselves with plenty of AI output and nobody able to connect it to a sound commercial decision.” — Andrew Mackintosh, Cobalt Consulting; former GM, Amazon Ireland

Andrew also cautions against handing over too much control before there is a track record of the AI’s decisions holding up. The opportunity is real, but agencies and clients need to test where these tools are reliable.

The Amazon answer may not be the business answer

Ant Finch of Vendor Rocket puts the commercial challenge plainly: “Plenty of recommendations will be right for Amazon but potentially wrong for your business.” A high ROAS product might have a thin margin once freight, returns and trading terms are included. A replenishment recommendation might ignore cash constraints, minimum order quantities or lead times. Matching a price drop to protect the Buy Box could weaken pricing across other channels or damage a more valuable retail relationship. Those are decisions an agency cannot make well from an Amazon dashboard alone. It needs to understand the client’s objectives, the economics behind the products and the consequences beyond the marketplace. As Ant argues, automation makes that judgement and experience more valuable.

What should agencies do now?

Take the services you currently sell and put them into three groups:

• Work a client or an AI tool will increasingly be able to do without you.

• Outputs that still need a skilled person to interpret and challenge.

• Commercial decisions for which the client wants informed advice and accountability.

The exercise is uncomfortable if much of your offer sits in the first group. It is also a useful starting point for changing it. Calling account management “strategic” will not bridge the gap. Agencies need to know which clients they serve best, how those clients make money, and where their own judgement changes an outcome. They need people and processes that can challenge a recommendation when the wider picture calls for it. Amazon’s tools will keep developing. Agencies should test what they can reliably do before changing how much control they give them. But the direction is clear: clients will be able to handle more routine work themselves. The agency opportunity is to become more valuable in the decisions that follow.

If your clients spent half as much time asking for reports and routine account updates, what would they ask you for instead?