Everyone has seen that OpenAI is launching a new checkout experience, which feels a little like affiliate marketing: product recommendations, a “publisher” that doesn't have an ad model yet, mentions of product feeds.
I understand how Affiliate LinkedIn is all in a tizz, this is our bread and butter, but we only have ourselves to blame.
Since the launch of GPT-3, we have seen a long line of legacy affiliate network execs putting out press releases about how excited they are about AI, and yet none of them have shipped anything meaningful.
Here is a long list of reasons that Affiliate LinkedIn shouldn’t concern themselves with Instant Checkout, and a few things we should do to drag affiliate marketing into 2025.
- If ChatGPT replaces your influencer, your influencer was useless
Firstly, we need to separate influencer marketing from product reviews.
Real influencers build trust and community; if GPT-5 can do their job, they never had an audience worth paying for.
- Top 20 blah blah blah
Now let’s take a look at “affiliate comparison sites”, which at this point are mainly large traditional publishers that have written a “top 10 laptop bags” page and put some affiliate links on it.
What are the chances they bought 50 lapbags and compared them? Well, judging from the commission they are making, almost zero.
They didn’t do real reviews, and they are not in your top 20 publishers anyway.
- ChatGPT reviews are also trash
And the funny thing is, that is mainly because they were built by scraping the web pages we as an industry created.
- The dirty secret: affiliate feeds are a mess
Have you tried to build a product on top of affiliate network product feeds? I have, it is a nightmare. Every network has its own “standard,” every brand within those networks uses different schemas and taxonomies, and the data itself is riddled with duplicates, broken images, inconsistent prices, and IDs that change constantly.
Instead of building real consumer products on top of affiliate network feeds, developers burn endless time writing parsers and fixers just to make feeds usable — which is why affiliate innovation has been stuck in the plumbing for years.
Legacy networks are awful at product feeds. There might have been a chance for affiliate marketing to power the agentic commerce revolution, but OpenAI took one look at the feeds networks provide and decided to do it alone.
- AI native discounts
Users love a bargain, that is why cashback and coupons have been the biggest publisher sectors for 15 years. That isn’t going to stop, and I don’t see Chat GPT doing a good job there.
A coupon agent? A personal cashback agent built by my bank, I can see that.
But we need new tools and new software that wasn’t built in 1999. This is the responsibility of everyone: networks, brands and publishers. If we don’t move with the times, we will all be looking for a new job.
- Real niche expertise still matters
Community and commerce have a genuine chance to work this time, as people go hunting for real recommendations. Real reviews by real people that you value the opinion of.
- The rise of long tail merchants
Instant checkout is all about the product, not the brand. This means that the long tail of merchants has an equal shot at getting the sale.
Affiliate marketing has always done a poor job in the long tail. There are too many manual processes for it to add up for them. These feel like fixable problems in 2025.
So, let’s stop shaking our fists at the clouds and actually modernise the technology that this industry runs on.
