You check your brand's AI visibility dashboard. Your competitors are showing up. You're not. Existential dread sets in: Do I need to create more content? Right now?
That instinct — panic, then produce — is exactly what got some brands in a never-ending loop over the past year. With no easy way to monitor visibility, some brands that wanted to compete were left shooting in the dark.
Robert MacCloy, CTO of AI search monitoring and measurement company Scrunch, outlined a stronger approach to building an AI-forward content strategy during a presentation at Cited, Stacker’s inaugural conference for brand journalism leaders. He also shared original research conducted by Stacker and Scrunch that showed why citations fade on AI results — and what actually makes them durable.
Instead of blindly producing content, here are five questions to ask before you do more — and what brands can do differently knowing what we know about AI visibility now.
Not as a knee-jerk reaction. The instinct, when a brand discovers it's invisible in AI results, is to find the gap and crank out content to fill it, then wait for instant citations. But that skips a step.
In his presentation at Cited, Robert suggested treating AI visibility like an experiment, not a content calendar. Creating new content without first monitoring your visibility is like a scientist starting without a hypothesis. Thankfully, AI monitoring is no longer a black box, and tools like Scrunch allow for quick diagnosis.
Monitoring takes some work, but it doesn't have to be hard, either. For brands that want to make a long-lasting impact on GEO, consider monitoring:
With all this monitoring data, look at which citations have outsized impact towards your goals, and work backward to cater to the domains that matter most for your brand.
LLMs do tend to cite Reddit as a credible source, but opting for a Reddit-first strategy can backfire, Robert said in his presentation at Cited. The problem is that Reddit doesn't carry the same credibility as other third-party sources, like the media. Reddit is often a source AI platforms use to decide what's worth pushing to the top of the internet, but it's rarely the dominant source for any industry, Scrunch’s research has found.
In finance, for example, NerdWallet and Bankrate are cited several times more often than Reddit, according to Scrunch. That gap between Reddit and high-authority outlets points to how AI systems weigh confidence signals before surfacing a recommendation. Earned coverage on third-party, high-authority outlets like industry mentions, comparison articles, digital PR, interviews, and expert contributions feeds directly into the credibility signals that AI looks for.
A brand can have well-optimized on-site content and still lose out to a competitor with weaker pages but stronger third-party pickup. This is how Stacker's earned distribution model helps — by getting brands placed in the outlets AI platforms already trust, rather than betting on forum volume.
None of this means Reddit is worthless. It can still matter when it's the decisive source shaping a specific customer's decision. But as a primary strategy, wide distribution to credible media outlets is the better bet for durability — and the one with a clearer line to actual citations.
As Robert said at Cited, “It pays to partner with the right publishers.”
There was a time when AI platforms didn't demote content made just for its own sake. And AI itself enabled brands panicking about showing up in GEO to quickly develop lots of content, without enough humans in the loop to improve it. It worked, for a while, until brands hit a cliff in web traffic.
Source: Lily Ray Substack
Within months of flooding the internet with formulaic content, platforms recognized the pattern: AI-bloated content wasn't adding value or meeting user intent, so the platforms began demoting it. Brands that earned short-term gains in AI visibility eventually saw their traffic fall below their pre-AI baseline. The “spray-and-pray” approach to content creation just doesn't pay off for long-term visibility, Robert said, based on what we’ve learned from the “half-life” of citations.
Stickier citations tend to rely on off-site validation, such as when a high-quality media outlet cites your brand. Monitor your presence, per citation, week by week to see which way you're trending.
Start with triage. Rather than starting from scratch, Robert recommended making a priority list of the content you want to maintain, seeing how it's performing in AI, and fixing pages that already exist but could be working better. Real value comes from answering the questions your audience has, then amplifying your point of view with third-party distribution.
New content earns its place when you're highly intentional about it, with original, specific, and differentiated content informed by the actual questions and tasks people bring to AI. That bar matters more now because AI searches aren't linear the way that SEO funnels were.
With AI agents, users jump between comparison, alternatives, and evaluation instead of moving in a straight line, so prioritize content that allows you to:
Content with staying power in AI citations is the kind humans actually care about — the kind that can't easily be replaced by a bot doing the writing and strategizing for you. AI tools, such as Scrunch's monitoring and Stacker's content recommendation tool, Sparks, can support your strategy. But having humans write the content matters more, not less, as LLMs work to mimic what people want.
If you've already been creating high-quality content, using a scalpel to make strategic fixes when testing your hypothesis will get you further than a sledgehammer.
Benjamin Chipman is a GenZ brand and content marketer at the intersection of media and storytelling. Informed by his experiences across the creator economy and the TODAY Show, Benjamin has a unique perspective about where traditional and new media converge. Passionate about where brands come together with culture and community, he brings this to all things brand and content at Stacker.
Featured Image Credit: Photo Illustration by Stacker // Canva