AI Shopping Brings Back an Old Affiliate Question: Who Gets Credit?

I've been training some folks on the basics of affiliate lately, and I keep coming back to a pretty common scenario...

Someone reads a thorough product review. A few days later, they ask an AI assistant to help compare a few options. They narrow it down to one, go to the retailer’s site and buy it.

So who influenced that sale, and who gets paid for it? AI shopping brings back an old affiliate question: Who Gets Credit?

If you’ve worked in affiliate marketing for a while, you know it’s not a new question. We’ve argued about attribution for as long as the channel has existed and never fully settled it. AI adds another wrinkle.

Picture a publisher that spends real time and money testing products and writing up a solid comparison. An AI assistant pulls from that coverage when a shopper asks for advice. The shopper gets what they need to make a choice, but they never land on the publisher’s site or click the affiliate link. The publisher may well have shaped the decision, and the affiliate report shows nothing.

I’m not saying that proves the publisher drove the sale. It’s still a good reason to ask whether the report is telling us everything.

I’ve spent years on the technology side of affiliate and running an agency, and one thing has held true the whole time: this channel is much better at recording what people did than explaining why they did it. A tracked click is real evidence. It’s just not the whole story.

There’s an easy answer going around right now, and I don’t really buy it. It goes something like this: sign up more content partners, call it your AI strategy, and wait for the growth. Before I’d recommend that to anyone, I’d want to know which partners, what kind of content, who it’s actually helping, what it costs, and what we’d need to see to keep funding it.

Paying for a placement doesn’t mean an AI tool will recommend you. Getting mentioned by an AI tool doesn’t mean anyone bought anything. And even when we can track a purchase, we still have to ask whether it would have happened anyway.

Brands have every right to expect accountability for what they spend. Publishers have every right to push back on a model that can ignore the work they put into helping someone decide. I think both sides have a point, and neither one should get to skip the hard questions.

So what should brands do right now?

I’d start with how your customers research your category. If people think hard before buying what you sell, reviews, comparisons, and specialist advice probably matter more than your reports suggest. Find out where that information comes from and whether the people producing it are part of your program. What makes sense will depend a lot on your product, your audience and how long your buying cycle runs.

Next, take an honest look at how you judge your partners. If everyone is measured only on last-click revenue, you’re probably missing some real contributions. On the other hand, if a partner tells you they drive discovery, they should be able to back that up. Your reporting and their pitch both deserve a hard look.

And when you test something, know why you’re testing it. If you’re funding an editorial partnership, agree up front on what you’re paying for and what a good result looks like. Watch the direct numbers, look at the supporting signals, and run controlled comparisons where you can. Be clear about what the results prove and what they don’t.

The payment model should match the work and the risk on each side. Sometimes a flat fee plus commission is the right call. Sometimes the math just doesn’t support it.

AI gives us one more reason to look hard at how discovery, influence and pay fit together. For brands, the job is to build a program around how customers really make decisions, then measure it carefully enough to know where the next dollar should go.

If you’ve been wondering whether your affiliate reporting shows what your partners actually contribute, I’d be glad to talk it through with you. The team at AIM can help you figure out what’s working, where the gaps are, and what’s worth testing next. Reach out anytime.