I wanted to start by saying that marketing on AI platforms is nothing new. Answer Engine Optimization (AEO), Generative Engine Optimization (GEO), AI SEO: we’ve had labels for this stuff for a while now. But let’s be honest. The field is still very new in the grand scheme of things. We’ve just been saturated in it so fast that it feels mature.
It’s become a massive part of digital presence for many companies, and we’re not used to having something this significant develop with so many blind spots in tracking, performance measurement, and best practices. It feels like there’s a new play nearly every month.
And as with any good platform category, advertising is entering the picture.
The AI advertising landscape
Ads in AI-powered experiences aren’t new. Google Search has already been pushing ads into AI-generated results and overviews. Perplexity tested ads, pulled them back. Some platforms, like Gemini, have taken a stand on not having ads at all. Though Gemini is Google’s, so AI ads already exist for the company.
Then there’s ChatGPT. Depending on which metrics you use, OpenAI still holds over 70% market share in the LLM platform category. That’s a lot in a world where it feels like everyone’s suddenly switching to Claude. And Sam Altman, who was one of the most outspoken voices against AI advertising, is now rolling out ads faster than anyone else. Money talks, Sam.
How OpenAI is rolling out ChatGPT Ads
OpenAI has been following the textbook platform ad rollout that we’ve seen for decades:
- Pilot the ad platform in limited tiers, limited markets, with big-spending advertisers
- Scale to more big spenders willing to pay on CPM
- Introduce click-based bidding and lower minimum budgets
- Bring a self-service platform to limited advertisers in limited markets
- Expand in phases with low minimum budgets across markets
The approach works well for validation. It’s no secret that the biggest spenders bring the majority of revenue, even when millions of SMEs are also advertising.
As I’m writing this, ChatGPT Ads are in phase four. Sponsored placements appear for Free and Go tier users, triggered by the context of the conversation rather than a single keyword. Early CPMs are reportedly in the $60 range, and CPC-based buying is becoming available. That means we’re approaching mass market, and it’s time for you to think about whether testing makes sense for your company.
The B2B digital marketing question: Should you even bother testing?
There are so many platforms to test all the time. I’m not asking whether you should commit to ChatGPT Ads as a permanent channel. I’m asking whether you should bother testing them at all.
Agencies like us, we’ll test them regardless. We need to understand how conversational AI advertising works and which clients could benefit. But for your B2B company, the calculus is different.
Three fundamentals of AI search advertising for B2B
With any passive platform like search, or in this case, AI-assisted discovery, you need to find out three things:
- Are people looking for things related to what you offer?
- Do people move toward a decision based on what they find?
- Can you realistically compete for attention in that space?
If you’re selling CRM software, the answers are probably yes across the board. Your buyers search, compare, and discover, and they’re usually ready to take action from that kind of channel. It’ll be competitive, sure, but there’s clear intent.
But if you’re selling deep-sea drilling equipment? The answer isn’t “no,” but it’s definitely not an immediate “yes.” You need to do more homework first.
This matters even more in AI advertising than in traditional search ads. ChatGPT ads are triggered by conversational context, not keywords. That means the platform needs to match your offering to what users are actually discussing. If nobody’s having conversations about your product category, there’s nothing for the ads to latch onto.
Check for life signs before you spend on AI ads
Before committing budget to ChatGPT Ads or any AI advertising platform, you want to know whether there are already signals worth amplifying.
Can you see traffic from AI platforms?
Remember: if you’re seeing traffic, the actual appearance in AI-generated answers and conversations is probably multiples higher. These platforms aren’t natural traffic drivers. Anyone clicking through to your site is going out of their way to do so.
Is the behavior commercially interesting?
Are visitors looking for general information to supplement a Wikipedia search, or are they showing buying signals based on their on-site behavior? There’s a big difference between someone researching what a category is and someone comparing vendors in that category.
How does your site perform in data fetching and AI crawling?
Are the platforms actually pulling information from your pages? You can dive into still-developing analytics tools to see whether your pages are being pulled and for what kinds of information.
When these signals are positive, testing becomes relevant.
How to test ChatGPT Ads without wasting your marketing budget
We’re often among the first to test self-service ad platforms for clients, and we frequently get early access. It’s usually a new world without established best practices, which means you need a proper approach.
The most important thing? Move on to testing and don’t debate it for weeks. I’ve seen it too many times: people try to find an answer by talking in meetings for weeks when it could be acquired by actual testing and a few hundred or thousand dollars. Just go for it.
Here’s what to keep in mind:
Limit the scope
Don’t try to do everything. Pick one product, one audience segment, one message. Remember that conversational ads work differently from search ads. Your copy needs to fit naturally into an AI conversation, not look like a banner ad.
Set a budget you’re willing to sacrifice
This is the cost of insight, not the cost of customer acquisition. Minimum campaign budgets can start from hundreds per month during the beta phase, with most advertisers allocating a few thousand for meaningful testing.
Manage expectations ruthlessly
You’re not going to explode your business success. You’re dipping your toe in the water to see if it’s worth wading in further.
Clarify the test focus
The goal is to validate the channel’s future potential. Nothing else. Yet.
When everyone’s on board and expectations are managed, give it a whirl.
Want to chat about whether ChatGPT Ads or AI advertising could work for your B2B brand? Get in touch!
Also read:
GEO Has a Measurement Problem You May Not Know About
GEO Is Not Killing SEO, It’s a Part of It