When Algorithms Compete: A Case Study in Cross-Platform AI Brand Visibility Challenges

When Algorithms Compete: A Case Study in Cross-Platform AI Brand Visibility Challenges

Alright, let’s cut to the chase. If you’re running a business that relies on being visible online (and let’s be honest, who isn’t these days?), you’ve got a new headache to deal with: competing AI platforms. It’s no longer enough to rank on Google’s search results or have a slick website. AI-driven responses from tools like ChatGPT, Google’s Bard (or whatever they’re calling it this week), and others are shaping how people find businesses, make decisions, and spend their cash. The kicker? These systems don’t agree. At all.

In my experience, the penny really dropped when a client of ours, a mid-sized e-commerce retailer, asked me, “Why does ChatGPT recommend us but not Google AI?” I didn’t have a good answer. Not right then anyway. That question sent us down the rabbit hole of cross-platform AI visibility. This article unpacks what we found, what it means, and, most importantly, what you can actually do about it.

The Problem: Diverging Algorithms, Diverging Results

Let’s start with what we know. Across the dataset we’ve built here at Contxt, 4,377 prompts tested across 45 different brands, the level of disagreement between AI platforms is shocking. Take ChatGPT and Google AI Overview, for example. They give conflicting answers 36% of the time when it comes to recommending brands. A third of the time. That’s like flipping a coin to decide whether your business gets mentioned or skipped entirely. And Perplexity? Its mention rate for the same brands is usually 20-30% lower than ChatGPT’s.

Why does this happen? Well, every AI platform is running on its own secret sauce (aka algorithms), pulling data from different sources, and ranking it differently. What one system sees as authoritative, another might completely ignore. The result? Brands can be highly visible on one platform and practically invisible on another. Across our network, some brands hit 100% visibility rates (shoutout to Red Fern, IPU, and Oak&More), while others are nowhere to be found. Poor DUO, for instance, lands a flat 0.0% on visibility. Zero. Related reading: Why 39% of AI Systems Disagree on Brand Recommendations (.... Related reading: this piece about the hidden gap: why brands are invisible to ai assistants.... Nada.

A Case Study in Visibility Extremes

Let’s dig into a real-world example to see how this plays out. One of the brands we work with, a small manufacturing firm called Castle Pumps, struggles with this exact issue. They operate in the marine and industrial pump space, a highly niche market. On ChatGPT, their mention rate sits at an abysmal 6.3%. On Google AI Overview? Pretty much the same story at 6.1%. But on Claude? It jumps to 28%. Still not great, sure, but four times better than ChatGPT. Why such a massive gap?

We dove into the specifics and found the problem: ChatGPT’s algorithms prioritise broader, more generalised sources of information. Castle Pumps, while technically accurate and useful, didn’t have enough backlinks or broader content footprint to register as “authoritative” in general searches. Claude, on the other hand, appears to weight niche expertise more heavily, which is why Castle does better there. Related reading: our article on why 37% of ai systems disagree on brand recommendations: ....

Contrast this with Red Fern, a manufacturing/industrial brand that nails 100% visibility across all platforms. What’s their secret? They’ve nailed the trifecta of content authority, consistency, and context alignment. They publish industry-specific content; they’re cited as a trusted source in their niche; and they’ve tested their prompts across platforms to ensure their language works everywhere. You can Contxt platform features to see for yourself. You can how AI visibility monitoring works to see for yourself.

What Actually Works: Lessons from the Testing Lab

So, how do you move from “invisible” to “everywhere”? Here’s what we’ve learned from the 58 experiments we’ve run in our testing lab. This bit’s important, so grab a brew and focus up.

1. First, understand where you’re missing

Step one: figure out your baseline. In our experiments, we ran 100 tracker prompts designed to assess brand mentions across platforms and industries. You can’t fix what you don’t measure, so ask yourself: where am I showing up? And where am I not? If you’re only testing visibility on one AI system (say ChatGPT), you’re shooting in the dark. Brands like Castle Pumps learned they were invisible on certain systems only after running prompts specifically designed to test AI platform divergence.

Our Contxt platform makes this process more systematic, but honestly, you can DIY with a spreadsheet and some persistence. Test a mix of buying-intent prompts (“Where can I buy marine pumps?”), recommendation-style queries, and even competitor-focused ones.

2. Tackle algorithm divergence head-on

The divergence between AI platforms is, frankly, a nightmare. Picture this: you optimise all your content for ChatGPT, but Perplexity or Claude ranks it entirely differently. This is the reality for 36% of prompts in our dataset. For example, one client found that Google AI prioritised local business directories more than ChatGPT did, so they leaned into optimising their presence there.

Look, it’s not rocket science, but you’ve got to customise your strategy for each platform. Test how queries are resolved, look at citation trends (i.e., which sources the systems favour), and adapt accordingly. In our buying-intent tests, we found significant variation across query types. A brand that dominated “best budget option” prompts might be absent for “luxury recommendation” ones. Language, intent, and context all matter massively.

3. Create platform-specific content

Here’s the truth: a one-size-fits-all SEO strategy doesn’t work for AI visibility. AI systems don’t just fetch content; they analyse and interpret it. We've covered this in more detail in AI Visibility Redefined: How Niche Brands Can Outshine In.... You might also want to check out our article on ai visibility metrics that matter: a practical framework .... This means you have to create content intentionally designed for how each system works.

Take Red Fern (100% visibility king). They produce highly specific, detailed guides for their sector, and they’ve tested these across multiple AI systems to see how well they “score”. On the flip side, Castle Pumps’ lower performance stems partly from relying on generic product descriptions that don’t demonstrate thought leadership or expertise.

If you’ve got product pages or blog posts, tweak them for AI consumption. Be explicit about what you offer, and don’t shy away from covering niche topics in depth. AI platforms love specificity.

4. Don’t forget ongoing monitoring

This one’s obvious but often ignored. Algorithms change. What worked six months ago might not work now. In our model-update-monitoring tests, we’ve seen performance swings of up to 15% for some brands after an AI system update. If you’re not tracking visibility regularly, you’re leaving a big blind spot.

Set up a cadence, monthly, quarterly, whatever suits. Run comparison tests across AI platforms to make sure you’re still being seen. And if you’re not, adjust. Treat this like ongoing SEO, because that’s basically what it is now.

Why This Matters More Than You Think

Some people might read this and think, “So what? It’s just AI. People still use Google, right?” Sure, they do. But AI-driven answers are fast becoming the default recommendation engine. It’s not just search engines, voice assistants, shopping bots, even business consulting platforms rely on generative AI to answer questions and recommend solutions. If you’re invisible there, those potential customers are never finding you. Simple as that.

The brands smashing it, Red Fern, IPU, Oak&More, haven’t stumbled into success. They’ve treated cross-platform AI visibility as a priority, not an afterthought. And at the end of the day, that’s what it takes. No shortcuts, just consistent alignment with how these systems work.

The Takeaway: How to Start Today

If all this feels a bit overwhelming, don’t worry. You’re not the only one. Start small, but start now. Here’s a quick game plan:

1. Assess your visibility: Run 10-20 prompts across ChatGPT, Google AI, and others. See where you stand.
2. Identify gaps: Are you being mentioned consistently? Are there platforms where you’re completely absent?
3. Analyse citations: Check which sources these systems are favouring. Are you listed? If not, why not?
4. Test intent-specific prompts: Does your visibility change based on customer intent (e.g., budget buyer versus premium buyer)? Adjust content accordingly.
5. Create AI-optimised content: Target key platforms and experiment with content formats, language, and depth.
6. Monitor regularly: Make visibility tracking part of your routine. Platforms evolve, so staying ahead means staying active.

At the end of the day, cross-platform AI visibility is a battle, but it’s one you can win with the right tools and approach. If you’re stuck or want to see how we run these experiments at scale, check out Contxt. It’s what we do, and we’d love to help.

Now, if you’ll excuse me, I’ve got to check whether my tomatoes have finally ripened at the allotment. Spoiler: they probably haven’t.