Change one page.
Move every AI answer.
Butterfly listens to what ChatGPT, Claude, Gemini and Perplexity tell your buyers about you, traces every wrong or missing claim to its source, ships the fix, and proves the answer changed. Every week, on autopilot.
First reading in 48 hours · No card · Cancel anytime
What's a good specialty coffee subscription in Canada?
Your buyers already asked. Something already answered.
An AI assistant is describing your brand right now, from whatever it found: an old pricing page, a retailer listing, a Reddit thread. That description is the only version most buyers will ever meet.
of traditional search volume Gartner expects to move to AI assistants by 2026
Gartner, Feb 2024weekly ChatGPT users asking questions a search box used to get
OpenAI, Oct 2025links on page two of an AI answer. You are in the paragraph or you are absent.
How answers workListen. Trace. Fix. Prove.
Visibility tools stop at a chart. Agencies stop at a deck. Butterfly runs the entire loop as software, every week, and only calls something done when the model says it differently.
- 01 Listen
Ask what your buyers ask
Every week Butterfly asks hundreds of real buyer questions across ChatGPT, Claude, Gemini, Perplexity and more, in every market you sell in.
- 02 Trace
Find the source of every claim
Each claim is extracted, scored against your Brand Truth file, and traced to the page, listing or thread that taught the model to say it.
- 03 Fix
Ship the correction, not a slide
Butterfly drafts llms.txt, JSON-LD, page rewrites and partner copy, then opens the pull request or ticket with proof attached.
- 04 Prove
Re-ask until the answer moves
Nothing is done until the model says it differently. The next reading verifies every fix and shows the before and after.
The files models actually read.
Not recommendations. Finished artifacts, validated before they ship, opened as pull requests or delivered as paste-ready copy, each one tied to the drift it corrects.
llms.txt
/llms.txtAllow, cite and reject directives, versioned per market, hosted or in your repo.
JSON-LD
schema.orgOrganization, Product, Offer, FAQ and Review schemas validated against schema.org.
Page rewrites
PR #88Topic, claim, proof and voice per page. Briefs that close a gap, not fill a calendar.
Answer blocks
FAQPageQuestion-shaped sections that give models a clean, citable sentence to lift.
Partner copy
PDPRetailer listings and marketplace descriptions aligned with your Brand Truth.
MCP server
mcp://Your brand attributes, products and proof, served to agents in your words.
One reading. Five minutes. Zero rewriting.
The weekly brief lands in Slack and your inbox: what changed in the answers, what Butterfly shipped, what was verified, and the one decision that needs a human this week.
- Answer score, share of answer and citation rate, with deltas
- Every fix with its verification result
- Regressions flagged with source and owner
- Board-ready, forwardable, no login required
Marlow Roasters, week 14
- Verified Vancouver, weekly shipping and $24 pricing now stated by 4 of 4 engines.
- Shipped Product JSON-LD for 12 SKUs merged (PR #91). Re-ask scheduled Thursday.
- Watch Gemini pairs you with Pilot Coffee as "larger alternative". Source: 2021 blog post. Fix drafted.
- Decide Perplexity cites a discontinued decaf. Retire the page or redirect? One click in Slack.
Published prices. Start in minutes.
Most GEO is sold as a retained engagement with a quote. Butterfly is software with a price on the page. Begin with a free scan and upgrade when the loop proves itself.
Annual billing: two months free. Full feature comparison
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.
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.