Guide · No smoke
How to choose an AI visibility provider without getting sold smoke
Before paying anyone to "show up in ChatGPT", ask these 10 questions. If they guarantee the #1 spot, claim special access to OpenAI or Google, or hand you a score with no confidence interval, it's smoke: inside an AI model there is no rankings table you can buy. What you can do is measure your visibility honestly and work the sources the AI cites.
The 10 questions
- Do you guarantee a spot or "#1" in ChatGPT?If they say yes, run. There is no spot to buy.
- Does one package make me show up the same across every engine?Each engine picks differently; a "yes" is a smoke signal.
- Do you have "special access" to OpenAI, Google or any LLM?Classic red flag. No such access exists.
- How many times do you query each prompt — confidence interval, or a single number?A single number with no range = false precision. AI is non-deterministic.
- Do you distinguish "fetched", "cited" and "mentioned"?If they blur them, they don't understand the mechanics.
- How do you connect visibility to business (or tell me when attribution is weak)?Honesty about the limits is a good sign.
- Does your plan include reviews or Reddit mentions that you plant?Astroturfing: banned by the FTC and bannable by platforms.
- Do you use hidden text or model instructions embedded in my site?Manipulation that no longer works reliably and risks penalties.
- Is your strategy to publish hundreds of pages a month?Scaled content abuse: penalized by Google, deduped by AI.
- Ask ChatGPT right now about THAT provider's category. Do they show up?Someone selling AI visibility who doesn't have it themselves tells you everything.
The most common smoke signals
- The "magic score". A precise number, no confidence interval, that swings every week.
- The impossible guarantee. "We guarantee #1 in ChatGPT" or "special access" to the model.
- Fabricated case studies. Round percentages with no period, no n, no verifiable client.
- Astroturfing. "We'll get you Reddit/forum mentions" or planted reviews.
- Manipulation tricks. Hidden text, cloaking or "instructions for the model".
- Content at scale. "Hundreds of articles/month" with no real authorship or expertise.
- Vanity metrics. Dashboards that only go up, with no n, period or interval.
What we do at Nethral
The opposite, by design: we measure your visibility engine by engine with a confidence interval (the trend, not a magic number), show you where the AI cites someone else in your place, and work the citable sources —with real editorial review, no tricks or fake reviews— so the answer includes you. And we tell you honestly when we don't know.