Pertento connects website experiments, visitor feedback and reusable learnings

Move from a suggested change to a test

Pertento combines website experimentation, surveys, analysis and feature flags. Its public page describes an AI editor that can build a proposed change on a live page, followed by the platform’s experiment workflow.

The useful distinction is between producing a variant and establishing that the variant helps. A simpler form or a different call to action is a hypothesis. The product’s demonstrations place those ideas alongside experiments and visitor feedback instead of treating an AI suggestion as a measured improvement by itself.

Keep the evidence beside the result

The platform advertises both frequentist and Bayesian statistics. Its voice-of-customer features connect written feedback with the variant a visitor saw, while a Learnings view looks across completed experiments for patterns and possible next tests.

The homepage includes a detailed sample dashboard with lifts, intervals and a fictional site context. Those examples are illustrative rather than a live assistant. They explain the interface and the kinds of questions it is designed to support; they are not Franklin results or a performance guarantee for another business.

Review, launch and learn deliberately

Pertento’s editor workflow asks users to describe a change and preview it before applying it. Feature flags provide a separate way to release a change without presenting that release as an experiment. This separation is helpful when deciding whether a team needs controlled evidence or simply an operational rollout.

For an initial evaluation, choose one measurable visitor task and confirm the tracking before interpreting a test. Inspect the proposed variant, compare it with the original and keep the test’s limitations attached to the conclusion. Use visitor comments to help explain an outcome, while remembering that comments and statistical estimates answer different questions. A convincing generated summary should make the underlying evidence easier to inspect, rather than replace that inspection.