Reviews the traffic
Reads how visitors move through each page, then picks the change most likely to move the conversions the owner named.
Feature · Autopilot
On Scale and Enterprise, Autopilot reads traffic, A/B tests a change, then publishes the winner. Hourly on Scale. Custom on Enterprise. Not on the free plan.
Scale and Enterprise only. Not on Starter or Growth.
Headline
Layout
Result
What it does
Autopilot runs on Scale and Enterprise. It never ships on a hunch. Every change is A/B tested against live traffic, and only the winners go live.
Reads how visitors move through each page, then picks the change most likely to move the conversions the owner named.
It never ships a guess. Every change runs as a real A/B test against live traffic before anything is published to everyone.
Once a test hits significance, the winning version rolls out. The owner can put a change back.
Keeps what works, then hunts for the next test. Hourly on Scale. Custom on Enterprise.
Reads the numbers
Autopilot starts from the same analytics the owner already has. It names what is working, what is slipping, and which change to try next. The figures in the panel are a sample mock.
Finding: the pricing page is the highest-traffic page but converts well below the home hero.
A/B testing
Headlines, layouts, and asks are tested against live traffic. Visitors are split. Autopilot waits for real significance before it ever calls a winner.
Statistical significance reached. Promoting Variant B to all traffic.
Winners
Once a test wins, Autopilot publishes that version. A pattern that fits more than one page can roll out there too. Every rollout stays in history, so Rewind can put a moment back.
sample · winning ask across 4 pages
What it weighs
Autopilot does not guess. Live signals from the site are weighed so it can pick the single change most likely to move the named conversion, then test only that one.
What it is for
Start on the free plan. Autopilot is the hands-off loop that runs after the site is already measuring, and only on those two plans.
Autopilot finds where visitors hesitate, then tests copy, layout, and asks until more of them convert.
A pattern that wins on one page can roll out to the pages it fits, so the site does not keep two versions of the same lesson.
On Scale and Enterprise, the loop does not wait for someone to open a dashboard. Growth Advisor still waits for a yes on Growth and Scale.
How it fits
Product
The history if a specific moment needs to come back.
Open →Product
Where visitors go, what they click, and what converts.
Open →Product
Suggestions to improve the site, accepted or declined.
Open →Product
Titles, sitemaps, and the signals search already reads.
Open →FAQ
Scale and Enterprise only. It is not on the free plan, Starter, or Growth. On those plans the owner still has analytics, and Growth Advisor on Growth and Scale suggests changes to accept or decline.
Hourly on Scale. Custom on Enterprise. Each pass still A/B tests a change against live traffic before a winner is published.
On Scale and Enterprise it can publish a winner after the test reaches significance. Every rollout is recorded and can be put back. Autopilot can also be switched off.
It splits traffic between variants and waits for statistical significance before calling a winner. It optimizes for the conversion the owner named, not vanity metrics.
Yes. Every Autopilot change is in the history. Rewind can return a specific moment. The owner stays in control of what stays live.
No. Growth Advisor suggests a change and waits for a yes. Autopilot A/B tests first, then publishes the winner on Scale and Enterprise. They are used together: suggestions on Growth and Scale, the hands-off loop only on Scale and Enterprise.
Sign up free. Analytics are built in. Autopilot A/B tests and publishes winners on Scale and Enterprise only.
Free forever plan · No credit card required