Google might still be the Goliath of search, clocking over 8.3 billion visits a month, while ChatGPT attracts a (comparatively modest) 1.8 billion+. But let’s be honest, what’s more revealing is the rate at which this gap is closing. The real picture is even closer when you add in Claude, Perplexity and other AI tools now used in B2B marketing research and buyer journeys.
The numbers don’t lie: AI search is here. And it’s rewriting the rules of visibility. But with change comes chaos. And opportunity.
Semantic content is now your best friend
In this AI-powered ecosystem, what matters isn’t just keywords—it’s meaning. If you’re not already optimising for semantic content and ensuring your pages are indexed by Bing (yes, Bing!), you’re already on the back foot. AI tools like ChatGPT and Microsoft Copilot increasingly rely on Bing’s index to fuel their answers.
So it’s not just about Google anymore. It’s about showing up where machines look for meaning.
What is really happening? The results of a real test
At Velo, we’re constantly exploring how to future-proof visibility for the B2B brands we serve. So, we ran a three-week experiment using a new monitoring tool called LLMRefs.com, a platform built to track brand citations and Share of Voice across AI search engines.
We focused on one keyword phrase “B2B marketing agency in UK”
Why? Because it’s a competitive space and we wanted to see:
- How much visibility changed in just three weeks
- Which platforms referenced which brands
- What we could learn from the volatility of AI search
And yes, we want to be found for this term too.
What we found: chaos, volatility, and huge swings
Over just 21 days, the landscape shifted wildly. Brands rose, fell, and sometimes disappeared completely from AI-generated search results.
A few standout stats:
- No brand stayed on top for more than a week
- One agency jumped from 5% to 48% Share of Voice in three weeks
- Several big names dropped off the radar completely, despite ranking high on Google
The takeaway? AI search visibility is highly volatile. What worked yesterday might not work today. And the traditional “set-and-forget” approach to SEO is officially obsolete.
Why this matters: you’re not chasing clicks anymore
In traditional SEO, the #1 position gets around 30% of clicks. But in AI search, there are no blue links—just direct answers. If your brand isn’t cited in those answers, you don’t exist in that context.
And unlike search engines, AI tools synthesise information. That means they’re choosing who to trust, reference and recommend.
You’re no longer optimising for search engines. You’re optimising for answer engines.
LLMRefs.com: What we learnt about the tool itself
The largest challenge is that without lots of manual searching, validating the results is difficult. However, it’s showing the kind of insight we would expect to see:
It’s built for AI-era visibility tracking.
Unlike traditional SEO tools like SEMrush or Ahrefs, LLMRefs focuses on tracking Share of Voice (SOV) in AI-generated responses—not just search rankings or backlinks. This makes it purpose-built for understanding how often your brand appears in AI search results, including platforms like ChatGPT, Claude, and Perplexity. We anticipate the goliaths like SEMRush will move into this space very quickly.
It reveals volatility in real time.
The tool surfaced massive weekly fluctuations in brand visibility—e.g., some brands vanished entirely from AI results from one week to the next, while others jumped by over 182% in SOV. This kind of insight is hard (if not impossible) to get from legacy SEO tools.
It highlights a new kind of insight: citation-based visibility.
Rather than tracking clicks or positions, it tracks how frequently and prominently your brand is cited in AI-generated answers. That’s a significant shift and helps marketers adapt their content strategies for AI-powered engines, not just human searchers.
It works across multiple models.
The analysis covered ChatGPT, Claude, Perplexity, and others, suggesting that LLMRefs provides multi-platform insights rather than being locked to one engine.
It’s early-stage, but promising.
While the tone of the write-up is measured, the use of the tool as part of a strategic test, rather than an afterthought, suggests it’s credible enough to shape decision-making. That said, no formal critique or side-by-side comparison with SEMrush is provided.
Recommendations: how to get ahead in AI search
Based on our observations, here’s what B2B marketers should be doing now:
Optimise for semantic richness
Create content that explores topics deeply and connects ideas. Think clusters, not keywords.
Monitor weekly, not monthly
Changes happen fast. Your current visibility might be gone by next Friday. Tools like LLMRefs are worth trialling.
Feed the Bing beast
Ensure your content is indexed by Bing, structured clearly, and offers value that AI engines can parse and cite.
Lead with thought leadership
AI tools reward original thinking, not recycled blog fodder. Invest in ownable perspectives, research, and provocative viewpoints.
Structure for machines
Use clear headings, schema markup, logical flow, and crisp language. If a human skims it and gets the point—so will the AI.
So what now?
AI search is a land of opportunity – but only for those who can adapt quickly. It’s not just about getting found, it’s about being chosen as a credible source by machines that don’t click, they decide.
You’re either in the answer… or invisible.