2025 Year in Review: AI + Affiliate + One-Person Company = 2026?

On the last day of 2025, a quick review of my year in affiliate. And an early Happy New Year to everyone.

My biggest takeaway this year: it’s not that affiliate has run out of opportunities, it’s that the way it’s played is changing.

1. AI Shakes Up the Affiliate Industry: Unease, but More to Look Forward To

If I think back five or six years to the first time I felt AI’s presence in affiliate, two things come to mind: one is Voluum’s Traffic Distribution (AI traffic distribution), the other is Kintura’s AI tracking, which came and went in a flash.

The principle isn’t complicated: learn from and analyze large amounts of tracking data, then automatically adjust the weights of Paths, Landers and Offers based on EPC or ROI.

These features were very useful in the A/B testing phase, and back then they genuinely saved me a lot of late nights.

Interestingly, even now, quite a few trackers still haven’t fully implemented this logic.

By 2025, AI is no longer just a feature inside a tracker; it has started entering every part of affiliate.

This year you could clearly feel more and more affiliates turning to social traffic. (See “Meta’s Andromeda Algorithm for 2026: Summary 1.0 (Part 1)”.) The reason is very practical: AI solved the creative production problem for us, and at the same time it forces you to understand the platform rules.

You can get by without knowing how to edit video or design, but you can’t skip learning how Meta and TikTok operate. When everyone is learning it, do you? If you don’t, it’s easy to get knocked out down the road. As I wrote in an earlier piece: “Why You Should Get into Media Buying Early”.

2. The Rise of the One-Person Company Owes a Lot to AI

Some people don’t much like the term one-person company; they feel it’s just solo entrepreneurship, more or less the same thing literally. But in my view there’s a difference in how you understand it.

A one-person company doesn’t literally mean one person doing all the work. A person’s energy is very limited.

Take me: I write for my WeChat Official Account and build an audience on X, I also run ads, tweak creatives and write copy, and on top of that I run a community, write tutorials and answer questions. Without AI’s help, that combination is very hard to keep up for long.

That’s also why more and more people keep bringing up the one-person company concept. Not because it sounds fancy, but because I think affiliate is inherently a great fit for the one-person company model: we don’t build products ourselves, and most of our time goes into testing, breaking down cases and looking for new plays.

And the underlying logic of affiliate is essentially arbitrage. As long as the arbitrage closes the loop, the system keeps running. For example, what I wrote recently: “ProPush Constructor: Build Your Own Survey Lander for Arbitrage”.

One link in that chain is PropellerAds’ Survey Exit traffic. Everyone knows Survey Exit is the exit traffic from surveys. When you see it, what’s your first instinct for how to play it?

I’ll write a separate piece later on how I run Survey Exit traffic. Consider that a teaser.

3. AI + Affiliate + One-Person Company Is How I See 2026

Put these three things together and that’s my call on 2026.

At least in the second half of 2025, it was already starting to work. When I first came back I wrote “2025 Push Publisher Starter Guide: Catch the Last Window of Opportunity”. Many of the plays that still run today fundamentally rely on AI’s help too.

While everyone started following along and testing “Facebook × Zeydoo: The Basic Playbook (Part 1)”, I also used AI to generate video creatives and tried more plays. So there will be another piece later: Facebook × Zeydoo video creative plays.

One more thing I want to say here: what we want isn’t simple copying, it’s fission. Through AI + personal experience + case studies + testing methods, find the path that fits you, and keep splitting off more plays along that path.

Finally, one sentence to close this year-end review, borrowing the Feynman Technique from Nobel physics laureate Richard Feynman:

If you can’t understand this article, it means I haven’t truly understood it yet. 2026, let’s keep going.