The world of online AI search is pushing forward quickly. Can radio keep up? Experts say the evolving world of search carries a lot of implications for radio sales and programming departments.
Online referrals for broadcasters are dwindling as audiences and advertisers increasingly let AI assistants to do the choosing. That presents an issue for radio since broadcasters can’t out-algorithm major technology platforms.
There’s been massive growth in information searches on the large language model platforms, so the impact on local media might not be immediately apparent, but it is only logical to think it’s coming.
Advertisers and agencies increasingly are using AI models such as ChatGPT, Perplexity, Claude and Gemini to brainstorm ad campaigns and how to spend advertising budgets.
That’s an issue for broadcast, according to radio thought leaders, since those digital helpers also seemingly give streaming platforms a built-in advantage. They say local media companies typically are poorly represented in generative searches.
“I don’t think bias is the right word for it, but I think there’s a couple of challenges for radio where AI-generated search is concerned,” said Dave Casper, senior VP of digital services for Radio Advertising Bureau.
“The big one is that radio is an audio business for the most part, and AI can’t see a lot of what radio does because radio is audio.”
LLMs know very little about local radio, Casper said, and they’re limited to what they can see on station websites and glean from other online sources. In many cases, that’s a relatively limited dataset.
RAB recently concluded a live presentation series designed to help broadcasters harness artificial intelligence to support their sales efforts and uncover opportunities.
AI is “rapidly reshaping how sellers work, research, prospect and communicate,” Casper said.
The webinar series focused on ways broadcasters can integrate AI tools into their sales strategies through demos, peer insights and success stories.
Casper said AI search engines are not trained to know much about any given station.
“In a lot of cases, [AI search] is defaulting over to Wikipedia. Well, heck, a lot of radio stations probably haven’t looked at their Wikipedia page since 2005,” he said. “Or it’s looking at some Chamber of Commerce article that was written 10 years ago.”
Casper said broadcasters must learn to provide information that AI can learn from on an ongoing basis.
“That’s true for our station brands just as it’s true for our clients. Having a marketing blog for your station is really important now because that’s how the LLM can learn about the different opportunities your station offers to clients in your market,” he said.
“That’s something that an LLM can consume and learn about and use as part of conversations that helps to solve the problems that potential advertising clients are asking.”
RAB even recommends using Claude or ChatGPT to build website pages that are optimized for AI to consume.
“LLMs can do a lot of this work. They can do a lot of the heavy lifting for us,” Casper said.
“A tool like Claude can code that for you; it doesn’t have to be hand-coded anymore.”
According to Gordon Borrell, president and CEO of Borrell Associates Inc., broadcasters must also develop the expertise to help clients retain visibility in the digital world, by getting discussions going.
“Losing visibility is on their minds, but you could also open the discussion about how they’re using AI for marketing purposes. Ask whether they’ve been using AI for marketing suggestions. Also make sure you’re going onto Claude, Gemini and ChatGPT to see what the LLMs are recommending,” Borrell said.
Also keep in mind that the LLMs are biased to the user, meaning you.
“So if you’ve been asking a lot of radio-related questions, your answers will likely be radio-biased,” Borrell pointed out.
Another secret of AI search, he says, is that it relies heavily on Reddit.
Borrell called Reddit “an interesting platform in that it’s not a bulletin board were you can slip in the door, pin something about yourself or your product and walk away. You have to participate, contribute, interact.”
The genius of AI search, he says, is that it thinks before it slops the results on a page.
“I’d enlist an expert — a company, preferably — in GEO, in generative engine optimization, and learn from them. You’ll need to rely heavily on structured information like Apple Business Connect and Google Business profiles, Yelp, Facebook and directories,” he said.
“Also look for authoritative sources, not blabby free-for-all bulletin boards.”
Chris Brunt, director of digital revenue generation and AI for Jacobs Media, said consumers using generative search also have an effect on website visits.
“Overall, traffic to media websites has declined significantly in 2026 as users increasingly get information from Google’s AI summaries rather than clicking through to the sites where the news or information originated,” Brunt said.
The Wall Street Journal recently reported that CNN and Business Insider experienced traffic declines of more than 30%. Other reports suggest that 68% of Google searches now result in zero clicks because users find what they need directly on the search results page.
Brunt says website AI optimization, which was not even a formal skill 18 months ago, is now a priority for many business owners.
“The starting point for radio stations is the same as it is for small and midsize businesses: making their mission, services and content clear and discoverable by both search engines and AI platforms.”
Losing visibility
Broadcasters must weigh “nuanced considerations” in order to appear in generative AI searches.
Jerry Del Colliano, publisher of Inside Music Media, said online publishers are measuring the shift in search since it affects traffic and revenue, but radio largely is not.
“Radio on the other hand is still often measuring Google searches,” Del Colliano said.
When it comes to asking for listening options, he says ChatGPT is likely to recommend Spotify playlists, YouTube music, SiriusXM or a national podcast before mentioning a local radio station.
And as AI becomes a new discovery layer, “broadcasters risk losing visibility before losing listeners,” Del Colliano believes.
He says that in order for radio to be found in local searches, stations need to program locally and flood websites with press releases and marketing suggestions.
“As it stands right now, who cares what regional or national radio personalities do, want, need or like? To be relevant on the first page of searches, radio needs to fix their financial tinkering and be in the front of mind for local listeners.”
In addition, he said, radio broadcasters have to do a much better job feeding AI useful information about the station.
“If I’m running a radio company, I am loading AI with example after example of success stories. Radio operators think AI is a magical mystery tour, but you have to give it things to skim.”
More info: SEO to GEO
Experts say to appear in AI searches, radio broadcasters must transition from traditional search engine optimization techniques to Generative Engine Optimization.
To do that, broadcasters need to structure their website content into machine-readable, question-based formats, publish transcripts of broadcast segments and establish strong local and brand authority signals across the web.
Here is a list of AI search optimization tools for broadcasters that were generated by a Gemini prompt:
Write for natural language: Frame website copy around conversational phrases and long-tail questions (e.g., “How do I listen to [Station Name] in [City]?” or “What country music show plays on Saturday mornings?”).
Lead with direct answers: Start articles, DJ bios and station guides with a concise 40- to 80-word summary that generative engines can lift and quote easily.
Use scannable elements: Break up text with clear question headers, bullet points and short summary blocks.
Build dedicated FAQ pages: Create robust FAQ sections targeting local events, streaming instructions, contest rules and ad rates.
Publish audio transcripts: Convert on-air interviews, local news segments and specialty shows into text transcripts so AI models can index spoken-audio content.
[Do you receive the Radio World SmartBrief newsletter each weekday morning? We invite you to sign up here.]
