August 2026 AI in Marketing Roundup: Cheaper AI Wins, ChatGPT Citation Clues & More AI From Google

August brought some interesting evidence that businesses don’t necessarily want the biggest, most powerful AI model available. They want the one that gets the job done at a price that makes sense.

We also got new data suggesting OpenAI’s publisher partnerships may have a measurable effect on which sources ChatGPT cites.

Meanwhile, social platforms continued figuring out what to do with the flood of AI-generated content appearing in users‘ feeds.

And, because no monthly AI roundup is complete without another batch of AI features from Google, marketers got some new tools inside Google Ads and Analytics that we’ll look at, too.

Here’s what happened in August.

Low-Cost AI Tools Reign Supreme

For the past few years, AI companies have largely competed on a familiar premise: build a smarter model, and people will want to use it.

But some new data suggests that equation may be getting more complicated.

The Financial Times reported that Anthropic’s newest and most powerful model, Fable 5, hasn’t exactly taken over among business users. Fable 5 accounted for only about 11% of customer spending on Anthropic’s models more than two months after launch, according to spending data from Ramp covering 70,000 companies.

If you’re wondering why, price appears to be a big part of it.

Older and less expensive models can already handle a lot of the work businesses need AI to do. That makes it harder to justify paying a premium for frontier-level performance when a cheaper model can write the email, summarize the document, analyze the spreadsheet or complete whatever other relatively ordinary task you throw at it.

Google seems to have noticed this broader issue as well.

The company announced new pricing and cost-control options for Gemini Enterprise in August, including pay-as-you-go pricing alongside its existing per-seat subscriptions. Businesses will also be able to set monthly spending limits, receive discounts for spending commitments and, eventually, even defer certain tasks to off-peak periods for steep discounts.

For marketers and organizations, that’s pretty good news and further evidence that the „best“ AI tool for a workflow doesn’t necessarily need to be the smartest model on the market. If a cheaper or faster model can reliably handle a task, why not just use that one?

Study Shows That OpenAI’s Partnered Publishers Get Cited More in ChatGPT

We don’t know much about how exactly AI companies‘ licensing agreements with publishers affect the sources their models surface. But in August, we got some new interesting data.

A joint study from Press Ranger and OtterlyAI analyzed 129.3 million citations across seven AI search platforms in June 2026 and compared them against confirmed licensing agreements between AI companies and publishers.

The headline finding: Pages from publishers with OpenAI licensing agreements received 48% more citations on ChatGPT than pages from publishers without one.

Licensed publishers averaged 10.2 ChatGPT citations per cited page, compared with 6.9 among publishers without agreements. And when researchers isolated publishers that had agreements exclusively with OpenAI, that gap grew wider, with licensed publishers enjoying about 112% more ChatGPT citations per page than unlicensed publishers.

That’s certainly interesting. But there are a few important caveats.

First, this is observational research. It shows a relationship between OpenAI partnerships and ChatGPT citations; it doesn’t necessarily tell us precisely why that relationship exists.

Second, having a licensing agreement wasn’t a universal AI visibility advantage. OpenAI was the only company in the study whose publisher deals demonstrated a clear citation advantage on its own platform. Google-licensed publishers actually appeared slightly less frequently in Google AI Overviews than comparable unlicensed publishers, while Perplexity’s partnered publishers were roughly even with non-partners.

Perhaps the most useful finding for marketers, though, has nothing to do with licensing agreements.

Across 16 U.S. industries, niche and trade publications received the majority of news citations in 15 of them. Collectively, those publications received 213% more AI citations than mainstream media.

That’s a pretty compelling argument for taking digital PR seriously as part of an AI visibility strategy.

You probably aren’t negotiating a content licensing agreement with OpenAI anytime soon. But earning mentions, quotes and coverage in authoritative publications within your industry? That’s considerably more achievable.

And if AI platforms are already leaning on those sources when generating answers, that’s another reason to think beyond your own website when you’re working on AI visibility.

Google Adds More AI to Ads and Analytics

Google also spent August putting more AI directly into marketers‘ existing workflows.

On August 10, the company announced new AI and agentic features across Google Ads and Google Analytics.

Among the additions are AI-generated summaries designed to surface important account information, the ability to create visual reports using natural-language prompts and new benchmarking capabilities that let advertisers compare their performance against similar businesses.

Instead of manually digging through reports or building every visualization yourself, you can increasingly ask the platform to surface what matters or create what you need. The marketer still needs to understand the data, question the output and decide what to do next — AI just removes some of the steps required to get there.

Final Thoughts

August’s AI news felt a little more practical — especially with businesses paying closer attention to what AI actually costs and opting for cheaper models.

At the same time, we’re learning more about what influences visibility in AI search, while tools marketers already use every day continue adding AI features of their own. August, and even a few other months this year, have felt less about the next big model or breakthrough and more about how AI actually fits into the work — what it’s worth, where it helps and how marketers can get more from what’s already available.

Note: This article was originally published on contentmarketing.ai.