Is AI Search Actually Sending You Customers? Most Owners Have No Idea

Here is the honest answer: most businesses are partially blind to this, not completely blind. You can see referral traffic from ChatGPT, Gemini and Perplexity in your analytics when someone clicks through to your site. What you cannot see is the part that matters more, how often your business was recommended in the first place. A customer can ask an AI assistant for the best business like yours, get three names that are not you, and nothing in your analytics will ever tell you that you were left out. That is why measuring clicks is not enough. You also need to measure visibility. Here is how to check it yourself, what to record, and what to actually do about the results.

Why AI search barely shows up in Google Analytics

When someone asks ChatGPT "who are the best steel fabricators in Melbourne" or "best Indian restaurant near me", the assistant might recommend three businesses. If yours is one of them and the person clicks through, you may see a referral from the AI platform in your analytics. That is the visible slice.

Everything else is hidden. If your business is not mentioned at all, there is no signal anywhere telling you that you were considered and passed over. And if you are mentioned but the customer gets what they need straight from the answer without clicking, you will never see that interaction either. So your analytics can show almost nothing while AI is actively shaping who your potential customers shortlist.

This is why so many businesses are measuring the wrong thing. They are measuring clicks when they should also be measuring visibility.

If that feels familiar, it should. SEO went through exactly this shift. Years ago we measured success almost entirely by rankings. Then we started paying attention to impressions, click-through rates and visibility, because being seen mattered before anyone ever clicked. AI search is doing the same thing again. The question is no longer only "are people visiting my website", it is "am I appearing in the answers customers receive". The businesses that get ahead will not necessarily be the ones with the most AI traffic today. They will be the ones paying attention to their AI visibility before everyone else realises it has become a discovery channel.

The monthly AI visibility check you can run yourself

You do not need a tool to start. You need to actually run the checks, every month, across ChatGPT, Gemini and Perplexity, not just Google. Each model draws on different signals and presents information differently, so it is worth comparing all three.

Start with ten to fifteen questions, not dozens. The mistake most businesses make is only asking "who are the best [service] in [city]". That is one important query, but it is a single slice of the customer journey. Cover four types of question instead.

1. Category searches. The customer knows the type of business they want. "Best steel fabricator in Melbourne." "Best mortgage broker in Geelong." These tell you whether you are visible for your core category.

2. Problem-based searches. The customer has a problem but not a solution in mind. "Who can manufacture custom steel gates?" "My website gets traffic but no enquiries, who can help?" These are often the most valuable, because they are much closer to how people naturally ask AI.

3. Comparison searches. The customer is narrowing a shortlist. "Who is better for custom fabrication or off-the-shelf products?" "What is the difference between hiring a freelancer and a digital agency?" This is where AI starts influencing the actual purchase decision.

4. Trust and credibility searches. The customer is validating a choice. "Is this business reputable?" "What are customers saying about them?" "Why would I choose them over competitors?" These reveal what AI understands about your reputation, expertise and what makes you different.

Run the same prompts across all three platforms so you can compare.

What to record so the check is actually useful

Do not just tick whether you appeared. Build a simple spreadsheet and track the same columns every month so you are measuring a trend, not a one-off snapshot.

• Did we appear? Yes or no.

• What position were we mentioned? First, second, third, or not at all.

• How were we described? Did the AI explain your services and what makes you different accurately, or was it generic?

• Was the information accurate? Check location, services, pricing, opening hours and specialties.

• Which competitors appeared consistently? The businesses that show up across multiple platforms are the ones worth studying.

• Were there citations or links? If the AI referenced your website or another source, note where it came from.

• Has anything changed since last month? Visibility is a trend, not a single result.

The goal is not to obsess over one month. It is to build a baseline. Just as you would watch your Google rankings or website traffic over time, AI visibility is becoming a metric that deserves regular attention. Track it consistently and you will notice not just when you are losing ground, but why, so you can adjust before competitors do.

What "good" looks like, and what invisible looks like

Appearing at all is the floor, not the goal. As you read your results, there is a rough ladder worth keeping in mind. Invisible means you are not mentioned for the questions your customers actually ask. Mentioned means you appear, but possibly low in the list or described vaguely. Described well means the AI explains who you help and what makes you different, accurately. Cited or linked means the AI is drawing on your own content as a source, which is the strongest position because it means your material is shaping the answer.

The useful insight is usually not just "did we appear", it is how you appear. Being mentioned third with a generic description is a very different problem from not being mentioned at all, and the two need different fixes.

How to go from invisible to recommended

If you are invisible or being described inaccurately, the first rule is do not try to game the AI. Focus on making your business easier to understand. The businesses that consistently get recommended have a few things in common, and none of them are AI-specific tricks.

They are crystal clear about what they do. Their website answers the simple questions: who they help, what services they offer, where they operate, and what makes them different. A surprising number of sites do not communicate this well.

They have strong proof. Reviews, case studies, testimonials, project examples, awards and genuine expertise. AI is trying to recommend businesses it has confidence in, so evidence matters.

They have consistent information everywhere. Your website, Google Business Profile, social media, directories and other mentions should all tell the same story. If one source calls you a web design agency, another a marketing consultancy, and another a software developer, you are handing AI ambiguity. A consistent digital footprint helps it understand and trust you.

They publish content that answers real customer questions. Not content written for search engines, content written for people. The more clearly you demonstrate expertise around the problems your customers are trying to solve, the more signals AI has to understand when you are relevant.

You will notice these are simply good digital foundations. AI is rewarding businesses that are easy to understand and easy to trust, which is the same thing that wins human customers. (It is the same clarity and proof work that decides whether ChatGPT recommends you in the first place.)

What to act on, and what to treat as noise

AI models change constantly, so do not overreact to every fluctuation. The skill is telling a signal from noise.

Treat it as a signal when you see a pattern. If all three major platforms consistently recommend the same competitors and never mention you, that is a signal. If they all describe your business incorrectly, that is a signal. If they consistently misunderstand what services you provide, that is a signal. And if your visibility improves over three or four consecutive months after you have strengthened your website and digital presence, that is a signal too, that the work is paying off.

Treat it as noise when it is a one-off. One platform giving a different answer on one afternoon, or rankings moving slightly from month to month, is not worth panicking over. These models evolve rapidly and their responses vary with phrasing, updates and context. The businesses that succeed are not the ones chasing every wobble. They are the ones tracking long-term trends and steadily making it easier for both customers and AI to understand who they are and why they are worth recommending.

What to do next

Treating AI visibility as a vague worry is how you lose ground quietly. Treating it as something you measure each month is how you get ahead of competitors who are not looking yet. Start this month: run ten to fifteen customer questions across ChatGPT, Gemini and Perplexity, record the results in a simple sheet, and you will have a baseline most of your competitors do not.

If the check shows you are invisible, described inaccurately, or losing consistently to the same competitors, the next step is understanding why, and that is what an AI Visibility Audit is for. It looks at how AI, search engines and customers actually understand your business, and identifies the clarity, proof and consistency gaps that are keeping you out of the answer.

Check if your business is AI visible. 

Speak To Us Now

FAQ

How do I know if AI search like ChatGPT is sending customers to my business?

Partly through your analytics and partly by checking manually. Referral traffic from ChatGPT, Gemini and Perplexity shows up when someone clicks through, but that misses the bigger question of whether you were recommended at all. The reliable way to find out is to ask each major AI the questions your customers would ask, every month, and record whether you appear, how you are described, and which competitors show up.

Why don't AI referrals show up properly in Google Analytics?

Because analytics only captures clicks. If an AI recommends three businesses and yours is not one of them, nothing is recorded. And if you are mentioned but the customer gets what they need from the answer without clicking, that interaction is invisible too. So your dashboard can look quiet while AI is actively shaping who customers shortlist.

How do I measure my AI search visibility without a tool?

Run a manual check each month. Ask ten to fifteen customer-style questions across ChatGPT, Gemini and Perplexity, covering category, problem-based, comparison and trust questions. Record whether you appear, your position, how accurately you are described, which competitors recur, and whether anything cited your site. Tracking the same questions over time builds a baseline you can act on.

How do I get my business recommended by AI assistants?

Make your business easy to understand and easy to trust. Be crystal clear about who you help and what you do, back it with real proof like reviews and case studies, keep your information consistent across your website, Google Business Profile, social media and directories, and publish content that answers real customer questions. These are good digital foundations, not AI tricks, and they are what AI rewards.

Should I worry if one AI tool doesn't mention my business?

Not on its own. AI models change constantly, so a single platform missing you one month, or rankings shifting slightly, is usually noise. Look for patterns instead. If all three platforms consistently recommend competitors and never you, or consistently get your details wrong, that is a real signal worth acting on.