AI Just Recommended Your Competitor Instead of You. Here's Why.

Ask ChatGPT to recommend a plumber in Adelaide. Or a commercial electrician. Or a financial planner.

It will give you three to five names. Confidently. With reasons. And your prospective customer will call the first one on the list, because a machine just told them it was the best option.

If your business wasn't on that list, you didn't lose a ranking. You lost a customer who never knew you existed. And according to SOCi's 2026 Local Visibility Index, ChatGPT recommends just 1.2% of local businesses. Perplexity recommends 7.4%. Compare that to the 35.9% visibility in Google's local three-pack. The bar is dramatically higher, the shortlist dramatically shorter, and the businesses making that cut are pulling away from everyone else.

The question worth asking isn't whether AI search matters yet. Twelve percent of Australians already name AI tools as their primary way to find information online, up from 5% a year earlier. Nearly half of Australians under 45 have used AI assistants to find local services. That number is growing faster than any channel shift we've seen since mobile overtook desktop.

The question worth asking is: what are the 1.2% doing that you're not?

The Traffic Nobody's Tracking Converts Better Than Everything in Your Dashboard

Before we get to the "how," let's talk about why this matters commercially.

AI referral traffic doesn't just send you a visitor. It sends you a visitor who's already been told you're good. The machine did the comparison shopping, weighed the options, and presented your business as the answer. That visitor arrives with a level of pre-built trust that your website normally spends five pages trying to create.

The data confirms exactly what you'd expect from that setup. Semrush's 2026 analysis found AI-driven visitors convert at 4.4 times the rate of standard organic traffic. RankScience reported a 14.2% conversion rate from AI search compared to 2.8% from traditional organic. Ahrefs found that AI search accounted for just 0.5% of their traffic but drove 12.1% of all signups, a 23x efficiency advantage.

These numbers should change how you allocate your attention. One hundred visits from ChatGPT are worth more than four hundred visits from a page-one Google ranking.

AI recommendation is this principle operating at scale. The machine functions as the most trusted referral source your business has ever had: perceived as objective, comprehensive, and expert. When ChatGPT says "here are the best options," it simultaneously delivers Cialdini's authority (expert endorsement), social proof (aggregated reputation), and what feels like independent verification. The visitor doesn't need your website to build trust. The trust arrived with them.

What AI Actually Looks At (Spoiler: It's Not Your SEO)

Here's where it gets uncomfortable for anyone who's spent years optimising meta titles and building backlinks.

• Relevancy (0.91): Does your business clearly match what the person is looking for?
• Brand mentions (0.87): How often is your business name mentioned across the web by other people?
• Reviews (0.61): What do your customers say about you on review platforms?
• Authority (0.52): Are you recognised as a credible source in your field?

The pattern is unmistakable. AI doesn't read your website and decide you're good. It reads what everyone else says about you and makes that judgment. As we've explored before, AI doesn't rank websites; it recommends brands. The data now shows us exactly which signals drive those recommendations.

Your Brand Has Mental Availability (or It Doesn't). Machines Can Tell.

Byron Sharp's research at the Ehrenberg-Bass Institute introduced the concept of mental availability: a brand's propensity to be noticed or come to mind in buying situations. Sharp demonstrated across 130+ brands in 13+ product categories that brands grow primarily by being easy to think of and easy to buy, not by being differentiated or deeply loved.

Here's what's remarkable about the AI recommendation data: it maps almost perfectly onto Sharp's framework, but for machines instead of humans.

When Sharp talks about mental availability for humans, he means the brand comes to mind across multiple buying situations because it's been linked to multiple Category Entry Points (the triggers that cause someone to think about a category). When AI "thinks about" your business, it does something structurally similar. It scans the web for signals that your brand is relevant, recognised, and talked about across multiple contexts.

A plumber who only exists on their own website has low mental availability for AI. A plumber who's mentioned in a local news article about emergency repairs, reviewed on Google and Yelp, listed in industry directories, featured in a supplier case study, and discussed in a Reddit thread about reliable tradies has high mental availability for AI. The machine encounters them across multiple "entry points" in its training data and indexed sources.

This is why brand mentions correlate so strongly with AI recommendations. Each mention is a data point that reinforces the brand's relevance and trustworthiness in the machine's model. Sharp would recognise this instantly: it's the same pattern he documented for human buyers, now playing out in algorithmic systems.

The practical difference? You can influence human mental availability through advertising (reach and frequency). You influence machine mental availability through earned presence: being talked about, reviewed, cited, and referenced by third parties across the web. Your business needs to exist in a thousand places online, not just one.

The Review Threshold That's Filtering You Out

Cialdini identified social proof as one of the most powerful drivers of human decision-making. His research showed that hotel guests reused towels 44% more often when told "most guests reuse their towels." The effect was even stronger when the proof was specific: "most guests who stayed in this room reuse their towels" boosted compliance by an additional 33%.

AI platforms have taken social proof and turned it into a binary filter. It's not a ranking signal that gives you a slight edge. It's a gate that either lets you through or locks you out entirely.

This explains something we've written about before: why customers look at your reviews instead of your website. Human buyers and AI systems are converging on the same trust signal. The difference is that humans might still click through to your site with a 3.8-star rating. ChatGPT won't mention your name.

For the average Australian SME, this has a clear implication. If you have fewer than 30 Google reviews or your rating sits below 4.3, you are invisible to the fastest-growing discovery channel in the country. No amount of SEO work will overcome that.

Why Optimising Your Website Alone Won't Get You Recommended

Here's the finding that should change how you think about digital marketing entirely.

Lily Ray's research found that when businesses create self-promotional "best of" listicles, Google's AI Overviews will cite the article as a source but omit the brand from the actual recommendation 69% of the time. You can be the source of the information without being the business that gets recommended. The AI uses your content, thanks you for nothing, and recommends someone else.

This aligns with broader research showing that 82 to 95% of AI citations come from earned media, not owned media. Your blog posts, your service pages, your carefully optimised landing pages: they are owned media. What AI trusts is what other people say about you.

This is also why Cyrus Shepard's analysis of 400+ websites found that proprietary assets (original data, unique research, tools, or resources that nobody else has) correlated at 0.357 with traffic growth. Sites with four or five of Shepard's five winning traits achieved win rates of 68-70%. Sites with none won just 13.5% of the time.

The businesses that show up in AI recommendations aren't just good at marketing. They're generating content that other people reference, earning reviews that machines trust, and building a presence across enough platforms that AI systems encounter them repeatedly. They're doing what Sharp would call building mental availability, but doing it for an audience that includes algorithms.

Meanwhile, SparkToro's 2026 research shows that only 276 out of every 1,000 Google searches now result in a click to the open web. That's down 26% in two years. The traffic funnel is narrowing. The businesses that thrive will be the ones visible in the answer itself, not just the ones ranking on a results page fewer people scroll through.

What This Means for Your Business

The shift from traditional SEO to AI visibility isn't a future scenario. It's happening now, and the data tells us exactly what to prioritise.

The businesses being recommended by AI right now aren't running some advanced technical strategy. They're doing the fundamentals of being a trustworthy, well-known business, and doing them consistently across enough platforms that machines notice. Sharp called it mental availability. Cialdini called it authority and social proof. In 2026, the machines are just measuring what good marketers have always known matters.

The difference is that now, the measurement is binary. You're either recommended or you're not. And the 1.2% who make the cut are capturing customers who convert at four to twenty times the rate of a standard Google click.

Your competitor might already be one of them.

Further Reading

• What Influences ChatGPT Recommendations (82-Factor Study): Neil Patel's comprehensive analysis of the factors driving AI recommendations, with correlation data for each signal
• An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied): Ahrefs' large-scale correlation study confirming brand mentions outweigh backlinks for AI visibility
• How to Rank in ChatGPT, Perplexity, and Google AI Overview: SOCi's Local Visibility Index with platform-specific recommendation thresholds
• In 2026, Less Than One Third of Google Searches Still Send a Click: SparkToro's latest zero-click research showing the narrowing traffic funnel
• Google Traffic Winners in 2026: 5 Data-Backed Traits: Coverage of Cyrus Shepard's analysis of the traits that predict traffic growth