The brands that figure out THIS feedback loop early: ads data informing content,
By Toddpie · February 10, 2026 · Curated by George's Blog
The brands that figure out THIS feedback loop early: ads data informing
content, content feeding discoverability, discoverability driving more
conversion data are going to have a real edge as agentic commerce picks up.
Been thinking about this after reading a Criteo engineering piece on how agentic recommendation systems work under the hood.
The gist: when AI systems recommend products, pure text matching (title,
description, bullet points) isn't what drives the best results. Models trained
on actual commerce data (clicks, add to carts, purchases) massively
outperform ones that just try to match words semantically.
Which got me thinking about a practical takeaway for brands running ads.
You already know which search terms convert. Your ad campaigns tell you
exactly what language people use right before they buy. But most brands treat that data as an ads-only insight. Meanwhile their PDPs are written in
brand-speak that doesn't reflect how real people actually search or talk about
the product.
If discovery is increasingly powered by models that learn from behavioral
signals (and this article makes a strong case that it is) then aligning your
PDP content with the language that actually precedes conversions is how you stay visible in a world where AI is doing the shopping.
This isn't about keyword stuffing. Nobody's saying paste your top 50 search
terms into your bullet points. But if "lightweight travel stroller that fits
in overhead bin" is what converts and your PDP says "compact infant mobility
solution" you're leaving signal on the table.