FAQ
Frequently Asked Questions
How the free audit works, what a GEO audit covers, how monitoring re-runs answer engines over time, and how agencies deliver our evidence reports to clients.
Basics
What Jincove is, who it serves, and how the output side of AI visibility differs from ontology and RAG.
Jincove is an AI-visibility (GEO) platform built by an independent studio. We audit and monitor what AI answer engines - ChatGPT, Perplexity, Gemini, Google AI Overview, AI Mode and Copilot - say about brands, and deliver evidence reports where every row opens back to a stored raw answer.
Agencies only: SEO, PR and brand agencies adding a GEO service line. You bring the client relationship; we bring the evidence pipeline - six answer engines, scheduled re-runs, and reports where every line opens back to a stored raw answer.
No. Ontology, RAG and grounding work on the input side: they feed your own agents your own business. We work the output side - checking whether external answer engines name a brand and cite its domain. We don't build ontologies, knowledge graphs or RAG pipelines.
Not yet. The data toolchain runs today and the web console is in early access. There is no instant online scanner - the way in is the free, human-run sample audit through the contact page, where you can also ask about early access.
Free audit
The human-run sample audit: what it covers, how fast we reply, and what happens after.
A human-run audit of ChatGPT, Perplexity and Gemini. Send one URL and one email - no card, no account, no sales call - and we reply within two business days with a mini report: what each engine answered, who it cited, and whether it named the brand.
Because it is run by hand. The free audit covers ChatGPT, Perplexity and Gemini; full six-engine coverage - adding Google AI Overview, AI Mode and Copilot - runs on our paid pipeline, starting with the Evidence Sprint from $199.
You get the mini report either way. If it surfaces gaps worth fixing, the usual next step is an Evidence Sprint (from $199): roughly 20 buyer questions across all six engines, a mention/citation split, a conversion gap and source gap list, and a prioritized fix list - every row opening to the raw answer. No obligation and no sales call.
Monitoring
Scheduled re-runs across six engines, diffs against your own timeline, and what we will never promise.
On a cadence from weekly - the default - down to every 15 minutes. Each run samples the full question set across all six engines and stores every answer verbatim, so any week's numbers can be opened back to what the engines actually said.
Each run is diffed against the brand's own answer history - not against a competitor, not against an industry average. AI answers drift across time and sampling, so we report a change only when it clears the sampling noise band; otherwise we stay quiet.
No, and we won't claim to. If an answer changed after a fix shipped, we show the timeline and the evidence - the stored answers before and after - without asserting causation. Turning a red row green is the fix's job; verifying whether it turned green is ours.
Method
Two booleans per answer, the gaps we derive from them, and why every number opens back to a raw answer.
Two booleans per sampled answer: mentioned - did the engine name the brand - and cited - did it cite the domain. From those two we derive the conversion gap (cited but never named) and the source gap (named while the citations point at someone else), with cited URLs stripped of UTM tracking.
We don't hand you a black-box composite score you have to trust. Any synthesized metric in a report - rates, splits, trends - opens back with one click to the raw answers it was computed from, with a stated sampling window.
Our own toolchain samples the six engines and normalizes each response into one record: the verbatim answer plus every cited source. Part of the real-time collection runs through a managed AI-scraper provider - we are downstream users of that data - while the records, analysis and reports are built by us.
The stored raw answer itself: verbatim text, every cited source, the engine, the timestamp and the sampling window - plus screenshots where captured. Every number in a report traces to those records, so your client can challenge any row and you can open it.
Agencies
Client workspaces, shared quota, white-label status, and how the data reaches your own stack.
Yes - an LLM audit is the natural entry point for an agency adding GEO. It tells you which buyer questions surface a client, whether the answers name them or only cite their content, and which sources the engines lean on - a baseline your SEO team can act on. The audit sets the baseline, monitoring re-runs the same questions on a schedule, and every report row opens back to a stored answer your client can check.
Each client gets an isolated workspace, so one client's questions, answers and history never mix with another's. Your agency draws from a shared quota pool across all workspaces - roster-friendly by design, since each client you add lowers your effective cost per client.
Full white-label - your logo, your domain, your report template - is coming soon. Today, you deliver Jincove-generated evidence reports inside your engagement, and we stay out of your client relationship.
Yes. Reports ship as working files - the sprint precedent delivered bilingual Excel - with every row traceable to the stored answer, so your team can rebuild the numbers in your own decks, dashboards and client formats.