Method

How Mynth works

Mynth checks how ChatGPT, Claude, Copilot, Gemini and Google AI Overviews talk about your company, finds out where their answers come from, fixes what’s missing and checks again every month. Here’s the whole journey, from your customer’s question to the next check.

This page describes a full setup. Scope and content are tailored to the brand and category. Last updated November 9, 2026.

Collage: a brass stopwatch and a steel caliper on cornflower blue, black and graph paper

In short

We ask AI every day

We ask AI every day

Every day, we ask ChatGPT, Claude, Copilot, Gemini and Google AI Overviews 200 to 300 of the questions your customers ask. That adds up to 30,000 to 45,000 answers a month. We see who gets named, and where AI found them. If a competitor shows up and you don’t, we fix it – and check again every month to see what changed.

200 to 300

200 to 300

customer questions a day

5

5

AIs we ask: ChatGPT, Claude, Copilot, Gemini and Google AI Overviews

30,000 to 45,000

30,000 to 45,000

AI answers read every month

Monthly

Monthly

report on what changed

The math

200 to 300 questions

×

5 AIs

×

30 days

=

30,000 to 45,000 answers per month

Steps 01–03

Measurement

First we look, then we fix. The first check becomes the starting point we compare everything against.

01

Your customers’ questions

We start with the questions real customers ask AI when they’re looking for someone like you: problems they want solved, options they’re comparing, questions right before they buy. We write them the way people actually talk, not in industry keywords. In a full setup, we ask 200 to 300 of them every day.

02

Five AIs

Every question goes to ChatGPT, Claude, Copilot, Gemini and Google AI Overviews. They answer differently and read different sources, so we check each one on its own.

For each AI, we note

Citations

Sentiment

Share of Voice

Visibility

03

Your competitors

We ask the same questions about your competitors. That shows how you come across – and who gets the spot when you don’t.

Steps 04–05

Understand

The next step is finding out why the answers look the way they do.

04

Where answers come from

citation graph

We map which websites AI points to in its answers, and which companies they lead to. That shows what actually shapes the answers in your field.

05

What’s missing

source-gap analysis

We compare the places that mention you with the places that mention your competitors, and list where they show up and you don’t. We sort the list so we start where it makes the biggest difference.

Steps 06–07

Fix

Everything we do fixes something we found.

06

Action plan

What’s missing becomes a concrete to-do list: listings to create or correct, journalists to contact, content to write and wrong facts to fix. We don’t do anything that doesn’t fix something we found.

07

Doing the work

Then we do the work: listings in directories and on trusted websites, contact with journalists, a clearer structure on your website, and content – or briefs for content you publish yourself.

Step 08

Prove

You see the effect in numbers, not assumptions.

08

Monthly report

Every month you get a report showing what changed since we started. We also check that your listings and details are still right. The report shapes next month’s plan.

Perception

Being visible isn’t enough if what’s said is wrong

We look at what AI actually says about you.

Sort

By topic

We sort what AI says by topic, so you see where your story is strong and where it’s weak.

Trace

To the source

When something AI says comes from a specific article or page, we show you which one.

Fix

At the source

If something wrong or negative comes from a specific source, it can be fixed right there. That’s why we find it first.

Illustrative example

What the first three months can look like

A Swedish B2B company with a technically complex service wants to be visible when buyers ask an AI about suppliers.

Month 1

Starting point

We write the questions buyers ask and put them to all five AIs, for the company and its competitors. That gives the starting point: how often they’re mentioned, how they compare with competitors, which pages AI points to and what it says. The action plan starts here too.

Month 2

Analysis and first actions

We map what AI says about the company, topic by topic, and where competitors get mentioned and the company doesn’t. The first listings and corrections go live, and articles start coming out.

Month 3

More work, new checks

The work continues and listings are checked again. We fine-tune the questions and find new things to fix. The report shows what changed for each AI and each topic, and the plan keeps rolling.

What Mynth doesn’t promise

What Mynth doesn’t promise

We can’t guarantee where you’ll appear in AI answers. AI writes every answer fresh, so no one can promise a spot. We tell you what we checked, what we changed and what happened – and we never report anything we didn’t actually see.

We can’t guarantee where you’ll appear in AI answers. AI writes every answer fresh, so no one can promise a spot. We tell you what we checked, what we changed and what happened – and we never report anything we didn’t actually see.

Frequently asked questions

Frequently asked questions with answers.

No. AI writes every answer fresh instead of showing a fixed list, so no one can promise a spot.

See how AI talks about you today

We’ll show you what AI says about your company right now.