In this article
The phrase “AI audience testing” can mean a lot of things. A generic website audit. An automated SEO scan. A tool that reads your copy and suggests edits.
Buyer Clone is none of those things. Here’s what it actually does.
Step one: building the audience panel
Before a single agent moves through your page, Buyer Clone needs to understand who that page is for.
You define your ICP: the industry, company size, job function, decision-making context, and buying criteria that characterise your target buyer. You describe what the page is trying to do — whether that’s a demo request, a signup, a download, or a direct purchase. You identify the products or services in scope, the key objections your audience tends to have, and what a successful conversion looks like.
From this, Buyer Clone constructs an audience panel. Not a generic set of personas, but agents shaped around the specific context of your offer and your market.
Step two: the simulated journey
Each agent approaches your page the way a real buyer in that role would.
They form a first impression from the above-the-fold content. They decide how much attention to spend — because real buyers don’t read everything, and neither do the agents. They move through the page in the order it presents itself, reacting to headlines, proof points, product claims, and CTAs the way someone with their background and priorities would.
At each stage, the agent is tracking:
| Signal | What’s measured |
|---|---|
| Comprehension | What did they actually understand about the offer? |
| Trust | What increased or decreased their confidence in each claim? |
| Objections | What questions came up that the page didn’t answer? |
| Attention | Where did they disengage, and what caused it? |
| Intent | Are they moving toward conversion, stalling, or preparing to leave? |
These signals combine into a journey map for each agent — a record of exactly where their reading slowed, stopped, or shifted direction.
Step three: the conversion brief
When each agent completes their journey, Buyer Clone synthesises the results into a conversion brief.
The brief doesn’t just list friction points — it explains what’s causing them, clusters them by theme, and suggests specific changes ranked by their likely impact across the most buyer types.
The friction themes are clustered: unclear value, weak proof, CTA hesitation, missing context, trust gaps. Within each theme, you see which agents triggered it, what specifically caused it, and what a fix might look like.
What makes it different from a generic audit
A generic AI website audit reads your copy and applies a checklist. It doesn’t know who your buyer is. It doesn’t know what your conversion goal is. It can’t tell you whether your technical evaluator is going to bounce at the pricing section because of how you’ve described your security model.
| Generic audit | Buyer Clone |
|---|---|
| Reads copy against a fixed checklist | Runs buyer-role-specific journeys through your page |
| One-size output regardless of market | Brief shaped entirely by your ICP and conversion goal |
| Suggests changes based on best practices | Suggests changes based on which buyer types are struggling and why |
| Doesn’t know your audience | Agents are built from your specific market context |
| No prioritisation | Changes ranked by impact across persona types |
Buyer Clone runs buyer-role-specific journeys through your specific page for your specific offer. The brief you get out reflects your market, not a generic optimisation playbook.
What it doesn’t do
Buyer Clone doesn’t change your page. It doesn’t implement suggestions. It doesn’t run A/B tests or generate copy.
What it does is give you a clear brief — a prioritised list of changes with reasoning — before your campaign goes live. What you do with that brief is up to you.
That’s by design. AI-generated changes to your marketing without your review is a different product entirely, and one we’re not building.