Why 39% of AI Systems Disagree on Brand Recommendations (and How to Fix It)

Why 39% of AI Systems Disagree on Brand Recommendations (and How to Fix It)

Here’s a question I hear a lot lately: why do AI systems recommend different brands for the same query? It’s a fair question. When you ask ChatGPT about the best running trainers, and then Google’s AI Overview the exact same thing, you’d expect consistent answers. But they’re not. According to data from our Contxt platform, 39% of prompts deliver conflicting recommendations between those two systems. That’s almost two in five queries producing wildly different commercial outcomes based purely on the AI system someone happens to use.

Think about what that means. Your brand might be highly visible on one platform and completely invisible on another. And if you’re not paying attention, that’s a whole chunk of your market slipping through the cracks. That gap, or what we call the “AI recommendation gap”, can be both a commercial risk and a massive opportunity. So, let’s break it down: why does this gap exist, what’s the real-world impact, and, crucially, what can you actually do about it?

The 39% Divergence: What’s Going On?

First, the brutal truth. There is no universal AI. Every system, whether it’s ChatGPT, Google AI Overview, Claude, or Perplexity, is trained differently, with different datasets, algorithms, and priorities. That’s why, across our dataset of 2981 prompts run through four major AI systems, we saw major discrepancies in brand mention rates. For example:

• ChatGPT had the highest mention rate at 75.8%.
• Google AI Overview lagged significantly behind at 46.5%.
• Claude and Perplexity pulled in at 67.6% and 55.0%, respectively.

Here’s the kicker: the same prompt, tested across all four, frequently delivered different recommendations. To put some numbers on it, 39% of prompts resulted in conflicting answers between ChatGPT and Google AI Overview. That’s huge! And we’re not just talking minor differences, like swapping one brand for another in a three-brand list. In many cases, one system would completely omit a brand the other heavily favoured.

Why? Two main reasons, as far as I see it:

1. Training datasets. ChatGPT and Claude are trained on vast but vague datasets. Google, on the other hand, leans heavily on its search index and real-time web crawling. If your brand isn’t optimised for the unique quirks of each AI’s training data, you’re playing a risky game.
2. Different weighting for user intent. Our “buying-intent” experiments confirmed this. For instance, a brand might show up when the prompt suggests high intent to purchase (“best running trainers to buy”) but disappear for exploratory queries (“how to choose running trainers”).

Simply put, AI systems interpret and weigh factors like relevance, authority, and intent differently. Related reading: our article on trust metrics decoded: how enterprises can evaluate the c.... Related reading: The Brand Invisibility Problem: Why Your Business is Invi.... That’s why no single strategy will cover all bases.

The Brand Visibility Winners (and Losers)

Another standout stat from our analysis of 32 tracked brands: visibility varies massively. Four brands sit in the “low visibility” bucket (below 30% mention rate), including TOCA Social (10.0%) and Contxt itself (13.0%). Ouch. On the flip side, 17 brands hit “high visibility” (above 70%), with a flawless 100% mention rate for Oak&More, Viking Pumps, and Specialist Blinds.

Why are some brands smashing it while others tank? Here’s my take.

Firstly, high-visibility brands often dominate within niche industries. Oak&More and Viking Pumps are great examples. If you’re the go-to brand in a tightly defined space, AI systems are much more likely to reference you. Related reading: our article on what is ai visibility and why your business needs it. For a deeper look, have a read of this piece about the cost of being overlooked: how decision-stage ai visib.... It’s all about being the loudest, most authoritative voice in your category.

But what about those floundering brands? From the prompts we’ve tested, it’s often a case of inconsistent digital signals. TOCA Social, for example, isn’t consistently well-represented across the sources AI systems rely on. Related reading: How to Optimize Your Website for Generative AI Systems in.... You can visibility tracking tools to see for yourself. And when your visibility is patchy, systems like Google AI, which weighs real-time web content heavily, are less likely to pick you up.

Why This Matters More Than You Think

So, what happens if you’re invisible to AI systems? Honestly, it’s not pretty. Let me give you a real-world scenario. A friend of mine runs a mid-sized ecommerce site specialising in sustainable homewares. Lovely stuff. But despite strong Google rankings and solid social campaigns, her brand barely surfaced in our AI visibility tests, just a 27% mention rate overall. Why? She wasn’t aligned with how these systems prioritise and process data.

The commercial fallout was stark. She started noticing a dip in organic traffic and, more alarmingly, a sharp decline in affiliate referrals (many of which now route through AI). You can the Contxt scanning process to see for yourself. People were asking AI systems for recommendations, and her brand wasn’t even in the mix.

Here’s the thing: if your competitors are visible and you’re not, you’re effectively invisible to a growing segment of the market. That’s a hard pill to swallow when you’ve poured time and money into building a strong brand elsewhere.

How to Close the AI Visibility Gap

Right, let’s get to the good stuff. What can you actually do to make sure your brand doesn’t disappear into the AI void? From what I’ve seen with our clients, the winning formula boils down to a mix of strategic alignment and consistent testing.

1. Understand the Rules of Every System

This might sound obvious, but it’s amazing how often brands don’t tailor their strategies for different AI systems. Here’s a quick cheat sheet:

• ChatGPT: Powerhouse for generalised knowledge, but prioritises brands with strong coverage in evergreen content.
• Google AI Overview: Heavily weighted towards real-time updates and high-ranking web content.
• Claude: A mixed bag, but often favours sources with clear authority signals like citations.
• Perplexity: Loves hyper-relevant, niche content.

If you want coverage across all four, you’ve got to balance authority-building, evergreen assets, and live signals.

2. Play the GEO (Generative Entity Optimisation) Game

GEO is the buzzword you’ll hear a lot in AI brand visibility circles. Essentially, it’s about feeding systems the clearest, most structured data possible about your brand. Sounds simple, but it’s a bit of a minefield in practice.

Here’s what works:

1. Clean up your digital footprint. Make sure brand mentions across the web are aligned, consistent, and regularly updated.
2. Focus on structured data. Schema markup, product feeds, and metadata are your best friends when it comes to helping AI systems understand and trust your brand.
3. Experiment constantly. Run tests to see how tweaks to your prompts, website, and content affect your visibility.

Of the 58 experiments we’ve run across 14 businesses, GEO-optimised brands consistently outperformed others by 15-20 percentage points in mention rates. That’s a huge jump.

3. Test, Analyse, Iterate

I know “test everything” sounds like vague advice, but stick with me. In our experiments, the brands doing well weren’t just throwing content out there and hoping for the best. They were actively testing and tweaking their strategies.

One of our best findings came from the 12 buying-intent experiments we conducted. We learned that brands often appeared in high-intent queries but vanished in exploratory ones. The fix? Creating content that bridges both intent types, providing depth for researchers and clarity for ready-to-buy customers.

Consider running your own mini-tests. Does your brand show up when you search for “best [product] for [use case]”? What about when the query is framed more loosely, like “how to choose a [product]”? If not, there’s work to do.

The Future of AI Brand Visibility

At the end of the day, the AI recommendation gap isn’t closing anytime soon. If anything, it’s likely to widen as systems evolve and specialise further. But that doesn’t mean you’re powerless. The brands that invest in understanding AI systems, treating them as a core part of their digital strategy, not a frustrating afterthought, will come out on top.

Here’s what I’d recommend:

1. Commit to regular visibility audits. Use platforms like Contxt to track how your brand performs across multiple AI systems.
2. Develop a GEO-first strategy. Prioritise structured data, establish authority signals, and test your prompts relentlessly.
3. Don’t assume your strong SEO game will automatically translate to AI. Align your content with the unique priorities of each system.

AI brand visibility might feel like a moving target, but the brands prepared to move with it will win the bigger prizes. The question is, are you one of them?

Ready to Take Control?

If you’re serious about bridging the AI recommendation gap, why not let us help? Sign up for a free visibility audit at Contxt. We’ll show you exactly where your brand stands, and how to fix any blind spots. Let’s make sure the next time someone asks an AI system for recommendations, your name’s the first to come up.