AI NativeWhich AI engines does AI Native measure?
AI Native measures four AI answer surfaces: ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews. These are the surfaces where buyers are actively asking questions and receiving AI-generated answers that name brands and make recommendations. Each surface works differently in terms of how it sources its answers, which makes measuring them separately informative rather than redundant.
How does ChatGPT answer differently from the other engines?
ChatGPT generates answers primarily from its training data, with an option to augment with web search. Answers tend to be discursive and confident in tone. The brand framing in a ChatGPT answer often reflects the accumulated web narrative as of the model's training, with web-search augmentation pulling in more current source material when enabled. This means parametric knowledge, what the model learned from historical training, can persist even when the current web narrative has shifted.
How does Gemini differ in how it answers?
Gemini draws on Google's search and knowledge infrastructure, which means its answers often reflect what Google has indexed and understands as authoritative. Citations in Gemini answers tend to point to sources that rank well in Google Search, so the overlap between your organic search presence and your Gemini citation set is often meaningful. Brands that have strong Google entity signals frequently see that reflected in Gemini framing.
What makes Perplexity different from the others?
Perplexity is explicitly a search-augmented answer engine. It retrieves current web pages, summarises them, and cites its sources inline. This makes the citation set in a Perplexity answer more transparent and more directly tied to what is currently ranking and cited on the web. A brand that earns mentions and citations on authoritative current sources tends to see that reflected quickly in Perplexity answers.
What are Google AI Overviews?
Google AI Overviews appear at the top of Google Search results for queries where Google decides an AI-generated summary is helpful. They are generated from sources that Google's search system considers relevant, and they appear in the context where most web-connected buyers still begin their research. AI Overviews can name brands, make comparisons, and recommend options, and they reach buyers who did not intend to use an AI assistant but encounter one in the normal course of searching.
What is AI Mode, and why is it separate?
AI Mode is a Google search surface that presents an extended AI-generated response with a conversational interface, as distinct from the standard AI Overview panel. It is opt-in for users and not yet the default experience for most queries. AI Native measures AI Mode as a separate, optional surface. Because it is slower to acquire and less universally available than AI Overviews, it is not included in the default engine set but can be enabled for products where coverage of this surface matters.
Do the engines give different results for the same prompt?
Yes. The same question asked on ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews can produce meaningfully different answers: different brand mentions, different recommendation ordering, and different framing. This is one of the reasons measuring each engine separately is important. A brand that leads in ChatGPT answers might be absent or listed without endorsement in Google AI Overviews for the same question. Each engine is its own measurement.
Does AI Native query the engines directly?
AI Native acquires answers through our data partners' infrastructure, which provides access to real responses from each engine. The measurement is of what the engine actually says, not a simulation or an approximation. For Google AI Overviews and AI Mode, the acquisition uses the search infrastructure rather than a direct API, reflecting how buyers actually encounter those surfaces.
Where does demand data come from?
Search volume and AI keyword demand signals are sourced through our data partners, not estimated internally. The platform queries real search volume data and AI-specific keyword signals for the head terms underlying your prompts, and uses the higher of the two values. AI Native does not fabricate or estimate these numbers.
Are the engines measured at the same time within a scan?
Yes. Within a live scan, prompt-engine pairs are measured concurrently up to a configured concurrency limit. Answers for the same prompt across different engines are acquired in the same scan window, so the results are comparable without the noise that would come from measuring engines days apart.
What happens if one engine fails to return a response during a scan?
If individual answer acquisitions fail, those answers are excluded from rate calculations and are not billed. The scan settles at the cost of answers successfully acquired. If an engine surface fails consistently across a scan, the scan summary flags the missing coverage so you know which engine results are incomplete. You are not charged for answers the platform could not retrieve.
Can I include or exclude specific engines for a product?
Yes. The engine surfaces included in a scan are configured per product. You can run scans against the full set of covered surfaces or narrow to the engines most relevant to your buyers. If your audience primarily uses one surface, focusing your scan budget there and expanding later is a reasonable approach.
Does the engine set change over time?
The covered surfaces will expand as new AI answer engines become significant for buyers. AI Native adds surfaces when they meet the threshold of being a real buyer-facing discovery path, not as a completeness exercise. When a new surface is added, you can opt into it for existing products without rebuilding your prompt set.