Alerts Ain’t What They Used to Be

Media alerting needs a deep rethink, providing more than just a clip notification.

The problem I have with alerts isn’t that they aren’t useful, they are, it just isn’t enough. News alerts were created by media monitoring vendors to replace human alerting workflows. The idea was that automated boolean and NLP capture applied against a source list would provide enough real-time information albeit with noise, to provide value.

When we released the Fullintel Hub AI Edition, one of our stated goals was to make people smarter faster. That meant, in large part, improving how alerts work and what information each can deliver. We know they can provide value but we also know according to Prowly, 43% of PR professionals cancelled paid tools in 2024.

Beat Reporting is Dead – Long Live Beat Reporting

This whole idea hinged upon something crucial: that media outlets produced news of distinct value, and that new digital sources would somehow do the same. What’s occurred is the growth of news syndication and a reduction in the number of beats that drove deep coverage. Take labor reporting as an example. In 2021, The Jacobin published a feature by Alex Press titled, “The Labor Beat is Dead – Long Live the Labor Beat”, in which they described how mainline publishers stopped covering labor as the movement declined.

Presse, in the article, highlighted that Chicago alone in the early 20th century had more than 72 newspapers and magazines that covered labor topics. That’s a lot of different voices covering a lot of different things.

Today the labor reporting pool is smaller, and newer journalists covering labour movements miss some of the context and history that backs crucial negotiations. 

“This can also look like shallowness: cheerleading union efforts rather than situating a fight in the balance of forces, historical precedents …” Presse wrote in the Jacobin.

So we have less news, and that news is less informed, and that’s only if the publisher has their own staff covering the issue. Most outlets will just source whatever AP wrote and move on.

At the same time, some of the highest-performing articles on Google are coming from the likes of NerdWallet, a credit card and loan evaluation website. It publishes a round-up of current strikes. According to Nerd Wallet, there are 109 different labor actions ongoing in the USA so far in 2025.

Where we get our information is changing, and how we format that information for consumption is changing too.

AI Overviews Growing in Use

We’re becoming more and more accustomed in our digital lives to having information spoon-fed to us after a prompt or a question. We published in our webinar with Ticket to Biotech in March that there were 60% fewer clicks on search where a Google Gemini AI overview was present. More recently, Seroundtable.com, a website that focuses on SEO research, showed that all major third-party research organizations like SimilarWeb, Ahrefs, and Amsive show declines in click-through rates of between 15% and 35% across most websites. This is attributed to AI overviews.

What that looks like for website managers, including those from news websites, is diverging lines between impressions (which Google tracks as eyeballs on search results) and actual clicks. Clicks are in blue. This is just one example that Seroundtable sourced, and it has many.

As communicators looking to affect dialogues in the media and the public, the shorthand of these overviews is a cheat code to deliver more information quickly. Other research by DemandSphere showed that 75% of the time AI Overviews source information from the top 12 Google search results. If you can rank in that top 12 for crucial keywords, or news you land in major news websites does the same your data should be present in those overviews. 

My bet is that as search behaviour changes from looking at articles and links to overviews and summaries, PR professionals will want both.

Links with Summaries – Summaries with Links

An AI overview on its own might not be trusted, it might be right, it might have deep data but people are aware that AI can produce hallucinations. What we’ve done in the our AI Overview and AI Alerts tooling is present both the links, and the summary of the issue in concert. You can look at the articles to validate the overview before forwarding it to your team or someone who needs to know.

Over time that might change, AI models might be more trustworthy, or preferences in information delivery may even become more personalized.

What are you looking for in your PR alerting tools, what would you like to see? 

If you are curious what that looks like, and how to begin using alerts as an early warning system follow the Fullintel Blog for more insights.