Asked often.
If your question is not here, email hello@butterfly.whyjs.com. A person answers within one business day.
What is Butterfly, in one line?
Butterfly is automated generative engine optimization: it listens to what AI assistants tell your buyers about your brand, traces each wrong or missing claim to its source, ships the fix, and proves the answer changed the following week.
What is GEO and how is it different from SEO?
SEO earns a position in a list of links. GEO (generative engine optimization) earns a place inside the answer itself, the paragraph ChatGPT, Claude, Gemini or Perplexity writes when a buyer asks a question. There is no page two in an answer. Either the model recommends you, describes you accurately, and cites you, or it does not.
Which engines do you cover?
ChatGPT, Claude, Gemini, Perplexity, Microsoft Copilot, Grok, Mistral and Meta AI. Google AI Overviews and AI Mode are tracked where available in your markets. Coverage depends on your plan; Starter covers four engines, Growth and Scale cover all of them.
Is this a dashboard or a service?
Neither alone. Butterfly is software that runs the full loop on a weekly cadence without waiting for a human: it asks, scores, traces, drafts and opens the pull request or ticket. You review in your own tools (GitHub, Slack, Linear, Jira). On the Scale plan a strategist also reviews each weekly brief before it ships.
How is Butterfly different from AYAN, Profound or Peec AI?
Visibility tools (Peec, Profound) tell you how often you are mentioned. Retained services (AYAN) add interpretation and hand-written briefs, priced as an engagement. Butterfly runs the whole loop as self-serve software with published pricing, from the first scan to the merged pull request to the verified re-ask. See the comparison page.
What does the free scan include?
25 buyer questions in your category, asked across three engines, with an answer score, the top drifts, each claim traced to its likely source, and one recommended fix. Delivered within 48 hours. No card required.
What is a Brand Truth file?
A versioned, machine-readable statement of what your brand actually is: positioning, products, prices, locations, proof points, claims you can substantiate and phrasings you reject. Every score is measured against it and every fix is drafted from it. It is yours, exportable at any time as Markdown, JSON-LD and llms.txt.
Do we need engineers?
No, but they will like it. Fixes arrive as pull requests against your repository or as CMS-ready copy with instructions. If you have no repository, Butterfly produces files your web team can paste in, and hosts your llms.txt and MCP endpoint for you.
How do you measure impact?
Three things, each with a before and after: answer score (accuracy of what is said), share of answer (how often you are recommended against competitors), and citation rate (how often your pages are the cited source). Every fix carries a verification date and a result.
Where is our data stored and is it used to train AI?
In Canada, on AWS ca-central-1, encrypted at rest and in transit. Your Brand Truth file, briefs and scores are never used to train models, ours or anyone else's. Model probing uses public assistant interfaces and APIs with no customer data in the prompt.
Can Butterfly run alongside our current tools?
Yes. Keep your SEO suite and your visibility dashboard. Butterfly ingests Google Search Console and GA4, posts to Slack, opens pull requests and fires signed webhooks, so it fits into the stack you already run.
How often do we get something?
A reading every Monday at 8:00 in your timezone: what changed, what shipped, what was verified, what is next. Fixes ship continuously as they are drafted. Urgent regressions (a wrong price, a competitor framed as your parent company) alert the same day.
Change one page. Move every answer.
Run a free scan and see what ChatGPT, Claude, Gemini and Perplexity tell your buyers today. First reading within 48 hours, no card needed.