Impermance is an AI visibility measurement and answer-engine optimization (AEO / GEO) practice based in Los Angeles. We measure how often AI assistants mention, cite and recommend a brand — with error bars — and do white-hat work on the brand's own pages and on the third-party sites those assistants cite.
Why “Impermance”
The name comes from impermanence. Ask an AI assistant the same question twice and you rarely get the same answer. That is the problem we measure: we ask every question many times and report intervals, not single scores.
How we work
Our method is public, and our research is pre-registered: the design and hypotheses are fixed before any answer is generated, and hypotheses that fail are reported as failed.
Today we sample OpenAI models with web search enabled; every report names the model that answered and the answer channel.