We collect fresh data from every major AI search engine daily, running your prompts in the specific market you've configured. The Adthena app is updated daily with data from the previous day. It can take up to 72 hours to show data for new prompts, as we load and index each prompt individually to provide the most accurate, real-time data.
We submit prompts exactly as you write them, with no rewording, no manipulation, and no enrichment. This is deliberate: it gives you full control over how specific or targeted the prompt is. If you want to name a persona, specify a location, reference a competitor, or phrase a query a particular way, that's exactly what the model sees. We run each prompt as a single turn, which mirrors how most users actually interact with AI search engines for discovery.
All collection happens in a clean, anonymous, logged-out state, with no accounts, no session history, and no personalization. This gives you an unbiased baseline of what a new user would see, rather than a view shaped by someone else's browsing history or profile. It also keeps results consistent and comparable across models and over time.
From each response, we extract three things: brand mentions, any outbound links the model provides, and the underlying sources the model cited when it searched the web. The same prompt set runs across all models on the same cadence so comparisons are like-for-like, and we aggregate across runs to give you a stable signal rather than a single noisy snapshot.
One nuance: some prompts don't trigger a web search, typically conversational or opinion-style ones, and when there are no sources, mentions, or links, there's nothing to capture. You'll see no data for those prompts on those days. If this affects your results, contact your Adthena account team to review your prompt set.