Dubbed “Blue” and “Orange”, the artificial intelligence tools will focus on backend processes and client interactions, respectively.
“Orange is the customer-facing large language model (LLM), and that’s on our website, also with agentic AI functionality for internal use,” said Wynand Beukes, chief digital officer at WeBuyCars, in a statement.
An AI agent is software that uses artificial intelligence to act autonomously and proactively on behalf of a user to achieve specific goals.
“Blue is a collection of machine-learning models that contains all the propriety information and pricing models of WeBuyCars, where we take a lot of factors into account, including our historic data, purchasing and selling history, market analytics, and trends. All those factors we use in pricing models, we update on a weekly basis,” Beukes said.
The Blue machine-learning software has already bought more than 2 800 cars “autonomously”, without any human pricing involved. “We’re scaling that up as we go,” he said.
WeBuyCars’ business model blends e-commerce with a physical presence and leverages data and technology to run its 17 “vehicle supermarkets” and roughly 100 “buying pods” nationwide.
“When we started, there was nothing. We didn’t have to deal with legacy, which was one of the great things,” said Beukes. “We wanted to be in control of the software and we wanted to be in control of the data.”
‘Fringe cases’
“When we get a new idea, we want to test it quickly using our experimentation platform. This allows us to expose the idea to a small segment of the market or employees, gather rapid feedback, and learn fast. If it fails, it fails fast – with minimal cost,” he said.
“If we can automate a certain percentage of our lead or buy lead management (where a customer wants to sell a vehicle to us), we can, with the same number of people, handle so many more leads. And then we let the petrolheads — the humans — focus on the fringe cases.”
Read: All the electric cars for sale in South Africa in 2025 – with prices
For example, there’s no reason why a human has to price a Volkswagen Polo Vivo, “because the volumes are so high”. But in fringe cases, like a 1974 Mercedes-Benz, “that’s a difficult thing to price. It needs understanding. So that’s where the human factor will always stay involved.” — © 2025 NewsCentral Media
Get breaking news from TechCentral on WhatsApp. Sign up here.