Why this conversation is worth having now
Most agencies still build their monthly reports around search rankings and social engagement, both still real, just no longer complete on their own, since a growing share of buyers now ask ChatGPT or Google's AI Overview before they open a traditional search engine. If your agency's report has no line for what AI engines say about you as part of your wider AI visibility, you are paying for half the picture.
The five questions to put in front of your agency
- Which engines do you actually test us against? ChatGPT, Gemini, Perplexity and Google AI Overviews behave differently; "we monitor AI search" without naming which ones is not an answer.
- What do you compare us against? You want named competitors, not a generic industry average.
- What is the one number you'll show me every month? Brand Mention Rate or an equivalent, tracked the same way each time, not a one-off screenshot.
- What will you actually fix, and in what order? Schema markup, directory consistency, content structure: vague promises of "optimization" are a warning sign.
- Do you test in Arabic as well as English, if that matters to us? AI tools answer differently in Arabic, and a GCC audience often searches in both languages.
What a credible answer sounds like
A capable agency names the engines, shows a baseline number before proposing new work, and describes the fixes in terms you could repeat to a colleague. If the answer stays at the level of "we handle SEO and AI is part of that," press further. SEO and AI visibility overlap but are not the same discipline, and treating them as identical is exactly how gaps go unnoticed for months at a time.
What to do if the answers are thin
You do not need to switch agencies just to get a straight answer. An independent AI Visibility Audit gives you your own baseline numbers, in both languages if that applies to you, that you can bring into the next conversation with your agency, or use on their own to judge whether the relationship is worth continuing as it stands.
Reading the answers once you have them
Write the five answers down and compare them against what your last three monthly reports actually contained. A wide gap between what the agency now claims to track and what has appeared in reporting so far tells you whether this was an oversight worth fixing together, or a pattern worth addressing directly at the next renewal. Either way, keep the written answers; they become the benchmark you check the agency against three months from now, once the fixes they promised should have had time to show up in your own results.