01What makes an A/B testing tool AI-native?
The AI does work you would otherwise do yourself: it decides which test to run, builds the variant, checks it, or reads the result. A chat box on top of an old editor is not the same thing.
- Choosing: which idea is likely to win, before anyone builds it.
- Building: variants from plain words, with a preview before launch.
- Working where you work: from ChatGPT or Claude, not only inside the tool.
- Reading: results explained in plain words, with revenue and profit.
02Which AI-native tool fits which job?
Apex by DRIP for choosing tests and working from ChatGPT or Claude; Intelligems for AI-built tests next to price and shipping tests on Shopify; Kameleoon for prompt-built tests and server-side flags in larger teams.
Disclosure
We build and sell Apex by DRIP. The other tools are described as their vendors document them, checked October 2026.
AI-native A/B testing tools by job (vendor-documented, October 2026)
| Job | Tool | What the AI does |
|---|---|---|
| Choose the tests that are likely to win | Apex by DRIP | A Likelihood to Win per idea from 4.3 million A/B tests across 151,000 shops; builds and reads tests from ChatGPT or Claude |
| AI-built tests next to price and shipping tests on Shopify | Intelligems | A building agent, a QA agent and an MCP server |
| Prompt-built tests and server-side flags in a large team | Kameleoon | PBX agents for ideas, building and analysis, an MCP server and a Claude connector |
More than twenty tools, sorted by what their AI does, are in our guide to AI in A/B testing tools.
