Skip to main content
Find it in Darwin Daily work → Experiments Open this screen ↗
Darwin tests different ways to start a conversation on eligible website pages. When a challenger earns a strong enough result, its opening can become the page’s new default. The next test starts from that improvement.

What changes automatically?

Experiments vary opening copy and eligible reply buttons using the page’s context. They do not rewrite your website, change your prices, or retrain the underlying AI model. Your knowledge, voice, guardrails, and locked reply buttons remain separate settings.

What does “better” mean?

Darwin evaluates recorded outcomes across the conversation funnel. Meetings matter, but the displayed expected-value lift is a weighted performance measure, not the same thing as a guaranteed percentage increase in booked meetings. A test needs enough traffic and meaningful outcomes before Darwin can make a reliable comparison. A challenger with a few early clicks is not automatically a winner. The app shows why it is waiting or why a result qualifies for promotion.

How hands-on do you want to be?

With a connected scheduler, automatic launch lets eligible pages enter testing and automatic promotion lets qualifying winners go live. Turn either off when you want to review that decision yourself. Choose your automation settings →
Experiments run on eligible pages within your plan and settings. Results and timing depend on traffic and visitor behavior; improvement is not guaranteed for every test.