How it works
The formula: AI-assisted conversion rate = orders from assisted shoppers ÷ number of assisted shoppers × 100.
If 2,000 shoppers talk to the assistant in a month and 160 of them buy, the rate is 8%.
Two rules change the result a lot:
- What counts as "assisted": a message sent, a click on a suggested question, a product added from the conversation. Each tool has its own definition.
- The attribution window: does the order have to happen in the same session, or within a few days?
Before comparing two numbers, check that they use the same rules.
Why it matters in e-commerce
This rate shows whether conversations turn into purchases. Split by page, category or question type, it tells you where the assistant helps most and where its answers need work. You can track it weekly, which makes it a good steering metric after launch.
Further reading: Fix Your Funnel Fast: Gen AI Secrets to Boost Conversion Rates
Real-life example: Kendra Scott
At Kendra Scott, the US jewelry brand, AI-assisted shoppers convert six times more than other shoppers. The brand wanted a new revenue channel. Cutting customer service costs wasn't the goal.
Common confusion
A rate six times higher doesn't mean the assistant multiplied sales by six. Shoppers who ask a question are often further along in their purchase already. To isolate what the assistant actually adds, you need an A/B test and a measure of incremental revenue.