How would you measure the success of OTOMOTO's AI car-search assistant in Poland?

Introduction

OTOMOTO is OLX Group's leading motors marketplace in Poland, and it has been one of the first OLX platforms to ship a conversational car-search assistant. Instead of setting twelve filters, a buyer can type "a reliable family estate under 60,000 złoty with low running costs" and get a curated shortlist; OLX has also made the assistant available inside ChatGPT. OLX Group reported investing about US$30 million in AI during its 2026 financial year, so leadership will want proof that assistants like this create value. The prompt asks how we would measure the success of OTOMOTO's AI car-search assistant. I'll cover context, goals, a North Star, supporting, guardrail, trade-off and counter metrics, and strategic initiatives.

Step 1

Product Context (3 minutes)

What it is: A conversational interface on OTOMOTO (and accessible via ChatGPT) that turns natural-language needs into searches, explains trade-offs between models, and hands the buyer to listings and dealer contact.

Users:

• Car buyers, especially those unsure which model fits their needs.
• Dealers and private sellers who receive the resulting leads.
• OTOMOTO's monetisation, which depends heavily on dealer subscriptions and paid listing visibility.

Value chain: Need → shortlist → listing views → contact (call, message, test-drive request) → purchase. The assistant acts at the top of this chain, so its value shows up downstream.

Business model link: Dealers pay for packages and visibility because OTOMOTO sends them leads. If the assistant produces better-matched leads, dealers see higher lead-to-sale rates and renew.

Assumptions: The assistant is live for logged-in and logged-out users in Poland, appears as an entry point on the home page and search results, and is in its first year.