Minimal changes
The critical fixes only, made inside your existing pages and design: missing facts stated clearly, headings corrected, the right internal links. Your website stays as close as possible to how it is today.
How it works
Every engagement follows the same evidence loop. You see the buyer questions, the answers each AI model gave and the pages behind them before you change anything — and you approve every change before it goes live.
Echo Tuner works in four steps. We audit how leading AI models answer the questions your buyers ask, fix the facts and sources those answers rely on, monitor each AI model as new versions change how they answer, and build genuine evidence about your business beyond your own website.
Step 1Audit
The audit starts with your website address and a conversation with you. It reads your site the way an AI assistant does, takes in what you tell us about your business, then tests the same questions on several AI models, for you and for your competitors.
Collect
Your public pages are collected and recorded with the date we saw them. We also take your input on your business and services — what you offer, who you serve, where you work and what sets you apart — so the audit reflects the business you run, not only what your website happens to say. Anything we could not read — blocked, broken or behind a login — is listed as unknown, never treated as missing content.
Profile
Your products or services, and the facts your pages state about each one, are extracted with the exact passage they come from. Nothing is assumed about your industry; the profile is built from your own words.
Questions
A shared panel of buyer questions for your market: broad discovery questions (“who offers…”), specific questions about one need (price, what is included, timing), and complex questions that combine several requirements. Brand questions are kept separate so they never inflate discovery results.
Page audit
For every offering we check whether your pages state the facts buyers need — how it is priced, what is included, who it suits, where you work, how to book, relevant qualifications, availability and more. Each requirement is marked supported, conditional, missing or unknown, with the passage or the pages we examined.
AI answers
Each question is asked to ChatGPT, Claude, Gemini, Perplexity and Grok with live web search — each assistant's everyday model in a Standard check, every current model in a Deep check. We keep each full answer and the sources it cited, then record whether each business was named, whether its own page was cited, whether it was presented favourably, and whether it was recommended when the buyer stated specific needs.
Compare
Your main competitors are assessed the same way, so you can see which facts they publish that you do not, and which of their pages the AI models cite instead of yours.
Off-site
Listings, reviews and independent sources that mention you are collected and checked: are they accurate, independent, and the kind of source AI answers cite?
Report
Findings for the business, each offering and each page, explained in plain language; the AI answers with their sources; the competitor comparison; and a prioritized plan. Everything is in your private client portal and can be exported.
Step 2Fix
We make the minimal, critical fixes that keep your website as close as possible to how it is today — or rebuild it as a GEO-optimized website. The plan ranks each change by how many buyer questions it affects and how much effort it takes, and every item links back to the evidence that justified it.
The critical fixes only, made inside your existing pages and design: missing facts stated clearly, headings corrected, the right internal links. Your website stays as close as possible to how it is today.
Your website rebuilt around the questions your customers ask: one clear page per offering, the facts AI needs on each, and a structure AI can read and cite.
Price basis, inclusions, service area, process and fit, stated plainly on the pages where buyers look; new pages where an offering has none.
schema.org markup for your business, services, products and FAQs — validated, and matching what each page visibly says.
You stay in control of every claim. We draft changes only from facts you have confirmed, show you each edit before it is published, record the page as it was before every change so it can be rolled back, and re-check the live page afterwards.
Step 3Monitor
AI providers release and update models constantly, and each one can answer differently. We continuously assess how new models behave and adjust your plan to improve your performance — while a frozen set of questions keeps every trend honest. You choose how often and how deep: monthly or quarterly, Standard or Deep.
Step 4Reach
Get other people talking about you — the evidence AI trusts most. Every service here is done in a format specifically designed for AI models to take in, so they learn about your business in a positive way from third parties — your customers, journalists and directories — not only from your own website. Always genuine: customers' own words and earned coverage, never fake reviews or paid links. Add any of these services, one at a time.
A simple way to collect genuine reviews from real customers, with replies drafted for you to approve — so AI hears specific, recent praise in your customers' own words.
One article a month that answers a real customer question in depth, written to be read and cited by AI models and published where it counts.
Earned coverage and expert quotes from the publications and "best of" lists AI cites for your kind of business.
Accurate, consistent business listings everywhere you appear — the details AI checks before it recommends you.
Always genuine. Reviews come from real customers and coverage is earned. We never write, buy or fake reviews, mentions or endorsements.
A typical first 90 days
A guide to the rhythm of the work, not a promised timeline for results. The order stays the same for every client; the scope depends on your business.
The audit itself takes about 1–3 business days: we confirm your business facts, agree the buyer questions and record the first measurement on each AI model before anything changes.
Publish the highest-priority improvements, with your approval, and verify each one is live and correctly marked up.
Repeat the frozen questions across several days, watch citations and inquiries, and note outside changes such as model updates.
Compare results with the baseline — including anything flat or worse — and decide together what to do next.
Next step
Create your free account in a minute — it starts nothing. When you're ready, choose a plan and pay at checkout; your results appear in your portal.