Each week we’ll gather headlines and tips to keep you current with how generative AI affects PR and the world at large. If you have ideas on how to improve the newsletter, let us know!
What You Should Know
What AI Search Means for the Future of Content
The future of search is AI, and it’s already starting to take hold. According to one projection based on 51,000 sites, traffic from ChatGPT has grown an average of 14.5% this year, while Google traffic dipped an average of 3%. Overall, Google still has a commanding lead with 41.13% of web traffic compared to ChatGPT’s 0.21%, but even the legacy search giant is adding AI Mode and Gemini to its Chrome browser.
As AI search continues to grow more popular, the strategies for getting found on the internet will change, too. News outlets are already pivoting away from reliance on organic search, as only one of the top 50 English-language news websites saw its August web traffic grow year over year. Many companies across industries are seeing the same declines in traffic, since AI Overviews and other AI models deliver more “zero-click” results, summaries instead of lists of blue links.
This can still work for many brands because the traffic from AI search tends to be much more valuable. Those users are more likely to convert because they arrive deeper in the sales process, having already gotten answers to many of their questions through AI.
So, what kind of content attracts AI bots?
The best answer depends on your industry, audience, products, and services. By digging into the data around the questions your prospects are asking, you can see when AI models draw on external sources, and which types of content are most cited and most influence the answers. (Reach out here if you want a demo on what this looks like for your company.)
Generally speaking, listicles are back, as are press releases. Original research, especially data that answers a common question or supports a common assertion in your industry, can be well-cited. How-to posts around problems you help solve or jobs to be done for your core audiences can earn a citation, as can case studies that show how you solved a particular problem for a particular type of customer.
While none of these are necessarily new, the approach should be totally different than traditional search engine optimization (SEO), which was predicated on keywords and antiquated 1,500-word count suggestions.
Optimizing content for AI tools is more about providing simple, concise, relevant, unique, and meaningful answers to questions. Instead of a common recipe page, where you have to scroll through the author’s life story before you get to an ingredient list, AI tools would prefer the recipe up top. Translated to business content, this might mean interviewing sales executives to learn what questions they typically get from prospects, then turning that into a compelling, well-structured blog post that answers what someone might be asking ChatGPT using proprietary data or expert interviews.
You can increase your chances of citation if you’re offering facts, data, or expert perspectives that no one else has, as long as they’re relevant to the searcher’s questions, which are longer and more customized than the typical Google search string. Short-sighted SEO strategies might suggest keyword stuffing to get “best running shoes” in your copy 10 times so you show up on the first page of search results, but if you share real testing notes about how shoes perform on city streets versus trails, you may be more likely to get picked up in an AI search for which running shoes are best for cross country.
There is no silver bullet to appearing in AI search because it’s still emerging, changing quickly, and many of the platforms’ algorithms are still black boxes. Start by looking at data around which companies are mentioned and which sources are cited for your space. That will give you the best sense of the content opportunities right now. And remember that your own content is only one source among many. Critically, you also need to increase your earned media initiatives (which Muck Rack says accounts for 89% of AI search results) to grow your AI search presence.
Elsewhere …
- PODCAST: How AI Can Use Brain Science to Become Your Personal Marketing Consultant
- Nvidia to Invest $100 Billion in OpenAI as AI Data Center Competition Intensifies
- Meet the New Recruit: Artificial Intelligence in College Football
- A Guide to AI’s Many Dazzling Futures
- Sam Altman Teases Mysterious But Expensive ChatGPT Features Coming Soon
Choose your thought time
What’s happening: Remember ChatGPT’s “model picker” that GPT-5 was supposed to make extinct? Well, now there are even more choices than before. In the upper-left corner, there are four options just within GPT-5: Auto, which decides how long to think on its own; Instant, which prioritizes speed; Thinking, which triggers reasoning models; and Pro, which thinks for a long time and generates “research-grade intelligence.”
How it works: If you select the Thinking model, and you go to type in a prompt, you’ll see there are two more options: standard and extended thinking. Standard balances thinking and speed, while extended thinks longer and is meant for more complex questions.
How to choose: Naturally, it depends on what you’re asking AI to do, but it’s hard to define clear swim lanes for each model. For any communications work, you probably don’t want an instant answer that may not be thoughtful. If you’re looking for analysis or working with data, you probably want to choose Thinking with extended thinking to trigger a reasoning model. If you don’t get the response you’re looking for, you can choose a different model within your chat and ask it to try again with that new model.
Quote of the Week
“Part of what makes them popular, in fact, is that they can feel more casually human, more like natural speech, than colons, semicolons and parentheses. Humans do not think or speak in sentences; we think and speak in thoughts, which interrupt and introduce and complicate one another in a neat little dance that creates larger, more complex ideas. (Or, sometimes, doesn’t: The copious dashing in J.D. Salinger dialogue is a great illustration of all the thoughts we leave unfinished.) This is the whole thing punctuation is for.”
— Nitsuh Abebe, author of a New York Times Magazine essay explaining why em-dashes aren’t AI tells after all
How Was This Newsletter?
