For years, digital marketing taught us to follow the click.
Someone searched, visited a website, and hopefully took an action. We could measure much of that journey and use those numbers to evaluate performance.
AI is changing that.
Today, someone can ask ChatGPT for a recommendation, receive a business name, and never click a link. They might search for that company on Google afterward. They might continue asking questions inside ChatGPT. Or they might abandon the search entirely.
From the business's perspective, those very different outcomes can look identical.
That's why I refuse to treat a missing click as proof of failure.
What I've seen firsthand
Several clients have found Monic AI Systems through ChatGPT.
Naturally, I want to understand how they found me. What question did they ask? What did ChatGPT say? Why did it recommend my business instead of another provider?
I've asked.
And often, they simply don't remember the exact words they used.
One prospect recalled looking for AI visibility companies that worked with businesses of approximately 15 to 20 people. That detail was useful because it told me something about the buyer's needs, even though I couldn't reconstruct the full conversation.
It also illustrated the attribution problem.
I knew the buyer had discovered my company through ChatGPT because they told me. But I couldn't reliably reproduce the exact path that led to the recommendation.
That's a meaningful distinction.
What I'm willing to measure
At Monic AI Systems, we monitor relevant buyer questions across AI systems and repeat prompts because the same question can produce different answers.
A business might appear in three out of three tests for one question, once in three tests for another, and not at all for a third.
Over time, those patterns help us understand where a business is appearing and how consistently.
But I don't interpret those results as proof that customers are purchasing.
I distinguish between being mentioned, being cited, being recommended, and ultimately being chosen.
Each tells us something different.
And each requires a different level of evidence.
What I won't promise
I won't tell a business owner that more mentions automatically mean more customers.
I won't claim a citation caused a sale simply because a company appeared in an AI-generated answer.
And I won't claim that an improvement in the questions we test represents the behavior of every person using ChatGPT or another AI system.
We can demonstrate changes in the visibility we measure. We can look for consistency and relevance. We can examine inquiries and customer-reported discovery sources.
But we should be transparent about what remains unproven.
That isn't a weakness in measurement. It's responsible interpretation.
Why I think businesses need more than another dashboard
Prospective clients have told me they don't need another dashboard.
I understand that.
Knowing where your business appears is useful, but the more important question is what you should do with that information.
Are you being recommended for the services you actually want to sell? Are AI systems accurately describing your expertise? Are your strongest buyer questions producing meaningful visibility?
And when they aren't, what needs to change?
A report tells you where you stand.
A useful strategy tells you what to improve next and why.
That's the standard I believe AI visibility providers should be held to.
Because the goal isn't simply to appear in more AI answers.
It's to become a business that AI can recommend for the right reasons, while remaining honest about whether those recommendations are translating into customers.


