This Week in B2B Signals: OpenAI Ads + GA4 Gets Worse For B2B Again + LinkedIn's New Predicative Tools + more!
Lets dive into what has been happening in B2B paid media this week, and where else to start but the launch of OpenAi advertising within ChatGPT
Let’s get into it.
OpenAI Ads: From Speculation to Structure
OpenAI formally rolled out its ad platform inside ChatGPT this week. I was discussing this on LinkedIn this week, and how it really launched with somthing of a whisper. However that colossal cost base isn’t going to get covered with twenty bucks a month subscriptions. Anthropic, and Claude, is rapidly pulling into the lead here and their stance is overtly anti-ads, as they broadcast to the world during the SuperBowl
As I posted on LinkedIn;
OpenAI quietly switched on ads in ChatGPT this week. Following their own guardrails they will not be in the middle of conversations. Not shaping responses. Appearing after chats, with safeguards around sensitive topics like health.
That alone signals something big: LLMs are no longer just tools. They’re media surfaces.
And the industry is splitting into two camps.
OpenAI: leaning into ads as a revenue layer to support massive model costs and scale distribution.
Anthropic (Claude): positioning against ads, even running a Super Bowl spot showing a rival chatbot serving creepy, intrusive placements.
Amazon (Alexa+): exploring commercialisation via ecosystem and commerce rather than pure media.
Google (Gemini): sitting on the world’s most sophisticated ad machine, but cautious about how conversational monetisation integrates with Search economics and as ever, being careful not to disrupt the search cash cow
Meta (AI assistants): likely to treat LLMs as extensions of its existing ad graph, not standalone media products.
Perplexity / others: experimenting with “answer sponsorships” and attribution-led placements.
Everyone’s solving the same equation:
- Models cost billions to train and run
- Usage is exploding
- Subscription revenue alone won’t cover it
- Ads are the obvious release valve
But conversational ads are not display ads with a new skin.
They force a rethink of the whole stack:
- Where does intent live when the interface is a dialogue?
- How do you insert commercial signals without breaking trust?
- What replaces impressions and clicks when the outcome is “an answer”?
- how do we handle the inherent uniqueness of each conversation, after all this isn’t 1/2/3 word searches but often big prompts with many users following up - all of which makes each users interaction highly personal
This is where B2B gets interesting.
Because B2B has always struggled with:
- low query volume
- long buying cycles
- invisible research (“dark funnel”)
- weak signal capture outside owned channels
LLMs change that. They become:
- research environments
- comparison engines
- vendor shortlists
- internal briefing tools
In other words: incredibly high-intent surfaces.
The winners won’t be whoever shoves banner-style ads into chat.
They’ll be the ones who:
- connect CRM and pipeline data to conversational contexts
- surface expertise, not promos
- align placements to buying-stage signals
- treat LLMs as decision infrastructure, not media inventory
The real question isn’t:
“Will chatbots run ads?”
It’s:
“What does advertising look like when the interface is thinking with you?”
We’re about to find out and the first version probably won’t be the one that sticks.
But whichever model nails this becomes the most valuable B2B demand surface since search.
Source:
Adweek – https://www.adweek.com/media/openai-launches-chatgpt-advertising-platform/
Digiday – https://digiday.com/media/openai-chatgpt-ads-rollout/
Key guardrails announced:
Ads appear after chats
Not in sensitive conversations
Not used to train model responses
Clear separation from output logic
Why this matters for B2B:
This is the first real indication that LLMs are becoming monetised surfaces, not just productivity tools.
We’ve talked about “intent forming inside AI workflows.”
Now that intent is about to become targetable.
The question for B2B isn’t just the obvious one:
“Can we run ads in ChatGPT?”
It’s:
“How do we measure signal when discovery happens before search?”
Signal is drifting upstream.
Google Tightens Consent & Modelled Conversions
Google quietly expanded enforcement around Consent Mode v2 in EEA markets.
Source:
Google Ads Help – https://support.google.com/google-ads/answer/13486246
Search Engine Land – https://searchengineland.com/google-consent-mode-v2-enforcement-2026-445002
Key change:
If advertisers are not passing required consent signals, modeled conversions will degrade significantly. This comes on the back of GA4 continuing to cause headaches across major B2B teams and agencies alike
For B2B:
Modelled measurement gets weaker, again - but frankly was it any good anyway?
First-party signal quality matters more
Server-side tagging becomes less optional
If you’re still optimising on platform conversions alone in B2B, your signal is about to get noisier.
The Trade Desk: UID2 Expansion + Retail Media Depth
The Trade Desk expanded UID2 integrations across additional CTV and retail media environments this week.
Source:
TTD Blog – https://www.thetradedesk.com/us/news
AdExchanger – https://www.adexchanger.com/data-driven-thinking/uid2-2026-expansion/
Two notable movements:
More retail media networks onboarding UID2
Greater CTV signal interoperability
For B2B:
Retail media pipes are maturing faster than B2B-specific pipes. Retail Media has attracted some of the sharpest minds in programmatic, and we - as B2B thinkers - can learn from what they are doing.
There’s a real question now:
Are B2B teams underutilising retail-style signal environments where business audiences overlap (tech, finance, professional services)?
CTV + UID-based targeting is no longer “experimental.”
Bombora Expands Intent Partnerships
Bombora announced new intent distribution partnerships expanding access across DSPs and CDPs.
Source:
Bombora Press – https://bombora.com/news/
MarTech Series – https://martechseries.com/sales-marketing/programmatic/bombora-expands-intent-distribution/
The interesting angle:
Intent is increasingly being packaged as modular, portable signal — not locked into single ABM platforms.
That aligns with where the market is going:
Best-of-breed signal coalition
Not monolithic ABM stack
The signal layer is fragmenting.
Distribution is decentralising.
I’m hearing rumours of a very interesting partnership announcement coming soon from Bombora, where they are going to be drastically upgrading Integrated ABM (iABM) in the TradeDesk via a specialist B2B partner
LinkedIn Adds Predictive AI Campaign Optimisation
LinkedIn rolled out expanded AI-powered bidding and creative recommendation features for B2B campaigns.
Source:
LinkedIn Marketing Blog – https://business.linkedin.com/marketing-solutions/blog
Marketing Brew – https://www.marketingbrew.com/stories/linkedin-ai-b2b-ads
Headline:
Platform-level automation deepens, and frankly, it is about time in LinkedIn
But here’s the tension:
LinkedIn optimises toward platform engagement metrics.
Most B2B businesses need pipeline metrics.
That gap is widening.
AI acceleration without revenue alignment = faster mislearning.
Programmatic Supply: Curation & Containerisation Momentum
This week saw further commentary on supply path optimisation, curated marketplaces, and containerised execution environments.
Source:
AdExchanger – https://www.adexchanger.com/programmatic/curation-2026-trends/
ExchangeWire – https://www.exchangewire.com/blog/2026-curation-shifts/
The shift:
Buyers consolidating supply paths
More curated, private, data-layered deals
Exchanges building smarter execution containers
For B2B:
Curation allows signal layering inside supply — not just targeting inside DSPs.
This is a structural unlock.
The future optimisation layer may sit closer to the exchange than the UI.
M&A: Data Infrastructure Keeps Consolidating
A few smaller but meaningful funding / acquisition moves this week in identity and data infra.
Crunchbase highlights:
https://www.crunchbase.com/
(Track: identity resolution startups + CTV measurement firms)
Pattern:
Money is flowing into:
Identity stitching
Privacy-safe matching
Clean room infrastructure
CTV analytics
Not into:
New banner formats
New ABM dashboards
Capital is betting on pipes, not panels.
That’s telling.
What This Week Actually Tells Us
Step back.
Across all these stories:
LLMs becoming surfaces
Consent tightening
Retail + CTV identity expansion
Intent decentralising
Curation strengthening
Infra funding rising
This wasn’t a “feature” week.
It was an infrastructure week.
And infrastructure weeks are the ones that change 12 months later.
The Macro Signal
Signals are moving:
Upstream (LLMs)
Closer to supply (curation)
Into retail & CTV environments
Away from cookie-based tracking
Toward portable identity
Meanwhile, most B2B teams are still:
Debating CTR
Running static TAL uploads
Optimising inside UI constraints
That delta is widening.
What Smart B2B Teams Should Be Doing
Audit where signal is forming before search
Reassess optimisation metrics vs revenue reality
Map where identity is portable vs trapped
Explore curated supply instead of open exchange sprawl
Invest in first-party HVA tracking before platform AI does it for you
Because platform AI will optimise.
The question is:
Optimise toward what?
If this edition was valuable:
Forward it.
Reply with what we missed.
Or tell me which of these threads you want unpacked next week.
Signals are drifting.
We’re here to track them early.
—
Want this weekly?
Subscribe here: