See what ChatGPT, Gemini and other AI assistants tell your customers about your brand.
More and more people ask AI to help them find products and services. Do you know what ChatGPT, Gemini and the other AI assistants tell your customers about you and your competitors? AIscopy finds out, with a diagnostic built around your business.
AIscopy takes its name from the -scopy in microscopy, a close look inside something. We look inside the answers AI assistants give your customer.
Launch price A$529 including GST. After the launch: A$659 including GST.
Mock-up answers with invented brands. Your First look uses real answers from the models behind ChatGPT and Gemini; larger and custom reports add Claude, Grok and others.
Anyone can type a question into ChatGPT. Three things about how assistants work mean one answer, or ten, will not show where you stand.
An assistant writes each answer fresh. Ask the same question twice and you get two different answers. If it names you in one answer in five, one check misses you four times out of five.
The different AI models behave differently and will prefer different features of your products. Moreover, the models are updated in the background, so last month's answer may not hold today.
Ask about your brand and the assistant describes it fluently. It describes a brand that does not exist just as fluently. From one answer, you can't tell which you're getting.
If your customer could ask an assistant which one to buy, or whether yours is any good, AIscopy shows what they are told.
answers in a single test, depending on how deep you go, every one asked as your customer.
Not sure? Ask for a proposal anyway. It costs nothing, and you pay only if you approve it.
Found, described and chosen, on two assistants, for one product and the customers you pick in your brief. It shows where you lose them, and to whom.
Coming after launch: a Full picture, which shows why you lose and tests two changes before you make them. By invitation: a Decision study, designed around a decision you have to make.
We read every brief before we start. Full refund if your brief can't work and can't be fixed, if we can't run your test, or if we're running late and you'd rather not wait.
We measure what the assistants say. No one controls it, so we can't promise it will change.
Which assistants do you test?
A First look tests the models behind ChatGPT and Gemini, asked directly through the access OpenAI and Google offer, so every question runs under the same conditions, many times. Larger and custom reports add the models behind Claude, Grok, DeepSeek and others. Every report names the versions and the dates.
Do enough of my customers use AI assistants for this to matter?
Six in ten Australian online shoppers use AI tools or features when they shop, at least sometimes, and 75% of those aged 18 to 39 (IAB Australia and Pureprofile, 2026). The report shows what they are told about you today.
What do you need from me?
A few minutes on the order form: your product and its page, what your customer would type to find it, and the customers you most want to win, in your own words. We send you a proposal within one working day: the questions we will ask, for whom, and what you get. You pay only if you approve it. If you are not sure a first look fits, see Is this for me.
Do I name my competitors?
No. You see the brands the assistant puts forward instead of you, including some you may not expect. Larger studies add a head-to-head against rivals you name.
What if a result is unclear?
Every figure comes with its range, and where the test can't separate two figures, the report says so plainly. If a check finds a fault in our measurement, we fix it before you see the report.
Is my report shared?
Your report, your brief and your name stay private. We never show them, or say you bought a report, without your written permission. Each report also adds the questions we asked and the assistants' answers to a pool, counted together with those from other reports so that no single report can be picked out. We use that pool for market reports and research. Your brand may appear there like any other brand the assistants name, never as our customer. We never use your brief to train an AI model.
How is this different from a monitoring tool?
Monitoring tools count mentions on a list of questions, daily or weekly, by subscription. We run a one-off test as the customers you pick, more than a thousand times, and show who is named instead of you and who gets picked when the assistant can name only one.
Can I run it again?
Yes. Order a new First look whenever something changes, and compare the new report with your last one.
Which markets?
Australia now, and other markets in a custom report.
Can I test a change, a second market or a bigger study?
Yes, in a Full picture, coming after launch: two changes tested before you make them, and up to three products or two markets. A Decision study, by invitation, is quoted for each decision. Compare the depths.
Can I ask a question before I order?
Yes. Email hello@aiscopy.com.au. We reply within one working day.