Run the diagnostic first
Before fixing anything, establish the facts: ask all four assistants your buyers' real questions, several phrasings each, and record who gets named and what sources the answers cite. If you are absent everywhere, the causes below apply globally. If you are absent from one assistant but present in others, the missing one's trusted sources are where to look. And if competitors' tools and firms show up while yours never does, their citations, visible in the answers, are literally the map to what they did that you have not.
Causes one to three: footprint, contradictions, content
One, no citable footprint: assistants corroborate before recommending, and a brand that barely exists in review platforms, directories, listicles and communities gives them nothing to verify. Two, contradictory facts: different pricing, categories or locations across your site, profiles and reviews make every engine hedge toward cleaner competitors. Three, content that describes instead of answers: pages about your passion for excellence contain nothing an engine can quote, while pages that resolve real questions in their first lines get lifted verbatim.
Causes four to six: category, staleness, density
Four, miscategorization: engines may have filed you under the wrong industry entirely, which you can test by asking each assistant directly what your company does; wrong answers trace to the sources that describe you wrongly. Five, stale signals: dated citations and untouched pages get discounted everywhere, and visibly so on Perplexity. Six, category density: in crowded markets the broad answer may be unwinnable this quarter, and the move is narrowing to the specific moments, niches and locations where you can plausibly be the best answer, then expanding from those wins.
The fix sequence, and how long it takes
Order matters because the fixes compound: correct contradictions and wrong facts first, usually days of work with effects in weeks. Publish direct answers for your top buying moments second. Then earn citations in the exact sources the assistants cite for your category, the slowest and most durable layer. Re-measure monthly across all four engines rather than reacting to single answers, since presence that survives rephrasing and re-runs is the only kind that counts. The free Aethon audit runs the full diagnostic for you, screenshots and citation trails included, and our ChatGPT-specific guide covers that engine's quirks in more depth.
