Free report: “5 Things $39.6M in B2B Programmatic Spend Taught Us.” Three years of managed spend and a billion tracked page loads, turned into five findings platforms won’t put in a pitch deck. Subscribe free to unlock it, plus every new issue of The B2B Stack.
Join 1,600+ B2B marketers, including readers at Dentsu, HubSpot, 6sense, Informa and DirecTV.
OpenAI just added shopping to ChatGPT.
You’ve probably seen the headlines “you can now buy on Etsy directly inside ChatGPT.” Both Etsy and Shopify will have buyable links presented in GPT in time for Christmas
It sounds like a classic consumer story.
It’s not. It’s actually a lot bigger then that in my opinion
This isn’t about checkout.
It’s about conversion.
If you’re in B2B you maybe glazed over the headlines but today we’ll explore how this could have enormous ramifications for tokenised intent within LLMs, throwing out new and powerful buying signals from the black box of zero click worlds. It got me quite excited! On the targeting side, segment vs signal is a good companion read.
The OpenAI commerce layer is not just a transactional convenience, it’s the first visible conversion protocol for the LLM era. The wider story is why AI didn't kill the dark funnel.
How AI Rewired the B2B Dark Funnel (and What to Measure Now)
Free report: “5 Things $39.6M in B2B Programmatic Spend Taught Us.” Three years of managed spend and a billion tracked page loads, turned into five findings platforms won’t put in a pitch deck. Subscribe free to unlock it, plus every new issue of The B2B Stack.
It merges search, discovery, and purchase inside the same interface — compressing the journey into a single moment of verified intent. For the basics, start with what B2B intent data is.
It tokenises behaviour.
It turns signals into currency.
And it has profound implications for B2B.
The Death of the Funnel, the Birth of the Protocol
For twenty years, marketing has built itself around funnel mechanics: attract, engage, capture, convert. TOFU, MOFU and BOFU. Acronyms and theory that never quite mapped in reality and were too simplistic at their core for the complexities of B2B
Each stage designed to measure quantifiable actions like clicks, views, form fills, demo requests. The ABM measurement guide defines what each stage should really be measuring.
But attribution on the open web is floundering. Cookies are gone, GA4 is broken, and buyers now operate in an invisible, fragmented world where most activity never hits your analytics. I dug into this in the last-touch attribution lie.
The Last-Touch Attribution Lie
Every marketing leader I've sat across from in fifteen years believes the same thing about attribution: that the model isn't lying to them, it's just incomplete. It’s missing data. Add more channels, stitch in more touchpoints, buy a better platform, and eventually the truth surfaces.
The LLM layer changes attribution for these reasons. I’ve heard mutterings of websites transforming into single page “original web” sites with a whole catalogue of machine indexable, structured content parked behind them but also large or even fully outside of the human eye. The back of house structured content sells the llm and the sexy front page sells the human. It also exposes how dated firmographic targeting can be.
This is all because of the seismic changes enacted by LLMs and how 1.8 billion people now engage with them.
These LLM’s don’t just observe behaviour, they now host it.
It knows what was asked, what was chosen, and what converted. It knows the history of chat too; it’s literally seen and chatted through growing intent with its human companions. It’s not a big stretch to think this tech will soon, if not today, know what you’re buying before you do
The conversation is the conversion. This layer hosts the intent data unlock of the century To see how to use that signal, look at reverse-mapping high-value actions.
That’s why this isn’t about a better checkout — it’s about a new kind of signal infrastructure.
When OpenAI, Google, and Perplexity start managing verified micro-transactions, they’re not just processing payments.
They’re building a closed-loop attribution engine - one that tokenises action and turns it into structured, portable intent data.
All in a llm walled garden. That’s the conversion protocol.
And B2B needs to see it for what it is.
This is one area that I don’t think B2B can afford to be 5 years behind on
Below is what this conversion protocol means for B2B signals and attribution, and how I’d get ready for it.


