Case Study

The Client

A retailer whose eshop visitors reached the product page but did not complete the purchase. Pre-sales questions — dimensions, compatibility, delivery time, returns — arrived by email and were answered hours later, once the customer had already gone.

Goals

Answer pre-sales questions at the moment they arise, and let the shopper examine the product well enough to decide without having to ask at all.

Challenges

Answers had to come from the real attributes and stock of each SKU, not from generic phrasing. The 3D viewer had to load fast on mobile, or it would hurt conversion rather than help it.

Our Solution

Chatbot grounded in catalogue data: Answers about attributes, availability and delivery come from the catalogue and stock, per SKU.

3D product viewer: Visitors rotate and zoom the product on the page itself, with models optimised to load quickly on mobile.

Behaviour-based suggestions: Related and complementary products are suggested from actual browsing, not a static list.

Hand-off to a person: Complex or post-sale requests are passed to the team along with the conversation history.

Implementation

We identified where in the journey visitors dropped off and which questions arrived most often by email.

We connected the chatbot to the catalogue and stock so answers are per SKU.

We produced 3D models for the main categories and optimised them for mobile.

We measured conversion with an A/B test before switching it on fully.

Results

Next Steps

Extending the 3D viewer to the remaining catalogue categories.

Adding order-status updates inside the same chatbot.