Carsales follows customers into AI chat apps

Carsales’ launch of an app inside of ChatGPT is connecting it to customers who are typically more advanced in their research or demonstrating high purchase intent, a sign of the potential payoffs of aligning with emerging AI ecosystems.

Once enabled, the app enables ChatGPT to “surface relevant carsales listings, including price and key highlights, with a direct path to the full listing or seller contact,” the marketplace said in an announcement in late March.

The company’s own consumer sentiment research identified the growing role of AI in car shopping, with 35 percent of buyers using the tools - mostly ChatGPT, with Gemini also emerging as influential.

Speaking to the iTnews Podcast, carsales’ executive general manager of technology and data Michelle Low said that AI offered a way “to rethink entirely how you can solve high friction points” in the purchase experience.

“Next to buying a home, buying a car is probably the highest consideration decision people can make. It's financial, it’s quite complex, it's emotional,” she said.

By having a presence inside of ChatGPT, Low said Carsales could “tap into emerging behaviours around how people do research” about potential car purchases.

“It’s important to be where [customers are] at, to offer assistance to the people who might have been on that platform [and are] really starting to think about their next purchase,” Low said.

“Historically, someone buying a car might've come onto the carsales website but not known where to start searching.

“Being on ChatGPT, where someone's starting a conversation means we can offer them the tooling that makes sense for where they are at in that conversation and help streamline that process a great deal by meeting them where they might be.

“That's the thing with AI: it affords us the means to solve these high friction or complex parts of the [purchase] experience completely differently than what we traditionally might have been used to in building web applications and traditional interfaces alone.”

On the conversion rates of engagements inside of AI to sales, Low said that the marketplace sees “much higher intent in terms of people that have come on [to carsales] from those [AI] interfaces.”

“We're also seeing that people, if anything, are better informed as to what they can get out of a trusted transactional marketplace like carsales,” Low said.

“So, without giving too much away, I will say it's been positive. It continues to evolve, and we're pretty excited at what we're seeing so far.”

Deriving value from data

Low initially began her time at carsales leading data, before adding technology responsibility to her remit as well.

“My team sits across product, platform, engineering, data AI, core applications, and the enterprise technology function,” Low said.

“It's quite a diverse group. The long and short of it is that my teams and I are responsible for everything the customer sees and interacts with on our platform.”

The technology and data strategy at carsales “is quite unified,” Low said, with an ultimate goal of optimising “moments that matter”, which she said made it “easier and smarter for customers buying and selling vehicles.”

From a consumer-facing perspective, the moments that matter are measured on impact.

“Ultimately, it is about impact - where a capability … moves the needle impactfully for the customer,” she said.

Carsales also has a significant B2B operation, working with car dealers and manufacturers. Again, the work of Technology and Data is around moments that matter, but there is an emphasis on co-designing appropriate solutions and innovations.

“We've had really deep relationships with our B2B customers for a long time, so a lot of how we think about product development is about co-thinking, co-designing, soliciting answers from them as to where their problems are,” she said.

“We find that when we co-design for the pain points, it's much better strike rate on being able to solve problems meaningfully for them.”

Low said that carsales had always been data driven.

“We genuinely actually build our products with data and treat our data as an asset,” she said.
Under Low, however, there has been a concerted effort to “shorten the distance between where data is produced and where it actually creates value”.

“Quite often that’s in two different parts of the business,” she said.

“We want to close the data-to-value distance pretty much anywhere in the buy-sell experience.”

This mindset is being brought to bear for B2B customers, with a relatively new dealer insights engine incorporated into the way they work with the carsales platform.

“A big part of how we assist [dealers and manufacturers] with data has always been about helping them understand things like inventory performance on our site: what vehicles are selling, what do they need to do to optimise sales performance of other vehicles, how can they move slow stock - things like that,” Low said.

“We call it Insights AI internally, but essentially it’s an insights engine that dealers can access, if they subscribe to it, to ask questions conversationally about current market dynamics with inventory,” she said.

Behind the scenes, the focus on shortening the data-to-value gap is being enabled through carsales’ enterprise data platform, and a focus on “thinking about [data] transport and architecture beyond the platform alone.”