AI CRO agent · Invite-only early access

Most AI hands you 50 test ideas. We hand you the few worth your traffic.

A CRO audit agent for ecommerce. Connect a URL and your GA4 — get a prioritised, evidence-backed A/B-test backlog where every hypothesis is grounded in real behaviour and stress-tested before you see it.

Behind a private beta with selected CRO agencies.
ContextCaptureBehaviourReasonFalsify
How an audit works

From your URL to a ranked, falsifiable backlog.

Seven stages with a shape: three evidence streams converge into one multimodal payload — then the model generates, attacks its own work, and ranks what survives. Tap any stage to see inside it.

Evidence in · three parallel streamsReasoning · narrowed to a ranked backlog
01 · Intake
Site & business context
3-step wizard5 ways to connectsets the tier
02 · Scrape
DOM, renders, funnel walk
~24 on-page signalsCWV + throttled-3G3 viewports
03 · Behaviour
GA4 funnel & drop-off
device CVR + channels+ Clarity signals
04 · Payload
Multimodal payload assembled
Three image blocks plus structured evidence, combined into one prompt.
3 images + textvision-reconciled
05 · Generation
Grounded hypotheses
Each tied to a behavioural mechanism and a citation; thin evidence is labelled, not inflated.
23 fieldsPXL 0–145 CRO pillarstwo-signal gate
06 · Falsification
Argued against itself
Failure mode, confounder, alternative, and the test that settles it.
focused audits · higher tiers
07 · Output
23-field hypotheses, ranked
Ordered by PXL, each with its tier, opportunity size and a tracking spec.
PXL-ranked5 tiersready to run
Tap any stage to see what happens inside it
into 01 · Intakefrom 07 · Output
Outcomes you log become priors for the next audit
A few minutes · end to endIt produces the backlog — it never runs your test

See the full pipeline, the payload and the tiers →

Why it's different

Three things most CRO tools won't do.

Anyone can generate a list of “best practices.” The work is deciding what's worth your traffic — and being honest about how sure we are.

Behavioural gatingThe two-sources rule

A structural finding isn't a hypothesis until behaviour confirms it.

We require two signals — a structural finding from the page and a behavioural pattern in your GA4 — before promoting anything to the backlog. The “small mobile CTA” isn't an issue until your mobile conversion rate says so too.

Adversarial falsificationThe second pass

Before you spend a test slot, the agent argues against itself.

On a focused, single-test audit, a second pass names the strongest failure mode, the likeliest confounder, the alternative explanation, and the single test that would tell them apart. You get the argument against the recommendation, not just the recommendation.

Portfolio learning RoadmapThe moat we're building

Lessons from one client inform the next — without sharing their data.

What works in furniture isn't what works in fashion. Cross-client patterns — abstracted to pillar and mechanism, never raw data — will inform future recommendations by vertical. Shipping in stages through 2026.

The honesty floor

Honest by design.

The brand rests on telling you what we don't know. These three behaviours are built into the agent — not bolted on.

01

The qualification gate

We tell you when your traffic can't reach significance — and refuse to fake confidence about it.

02

The confidence tier

The tier on the report reflects what we actually got from your data, not what you asked us to do.

03

The falsification pass

On a focused audit, we argue against our top recommendation before you see it — and tell you the one test that would prove us wrong.

Fit

Who it's for — and who it isn't.

The honest version. If you're in the right column, we'll tell you on the demo rather than take your money.

Built for

  • CRO agencies running active programmes across a portfolio of clients.
  • Ecommerce operators with 1,000+ monthly sessions and GA4 connected.
  • Teams running at least two A/B tests per month and tired of “best-practice” guess lists.
  • Brands where conversion rate is treated as a metric, not a vibe.

Not built for

  • Sites with under 1,000 monthly sessions. Not enough traffic to test anything honestly.
  • Teams without analytics. We'll tell you the audit is structural-tier and won't pretend otherwise.
  • Anyone looking for a “set it and forget it” SEO or marketing tool. This is a hypothesis engine, not a content factory.
  • Teams that want unmoderated AI recommendations they can ship without review.
Early access

See a real audit of your own site.

No generic demo. We run the agent on your URL and walk you through what it found.

15 minutes · a real audit on your site · no slide deck.