How Carousell’s CEO Says AI Is Turning the Marketplace Into a Safer Space

When you scroll through Carousell’s endless stream of second‑hand finds – a vintage lamp, a barely‑used iPhone, a hand‑stitched tote – you probably don’t think about the engine humming behind the scenes. Yet, according to the company’s chief executive, that engine is now largely powered by artificial intelligence, and it’s changing the way the marketplace works.

"We built Carousell to be a community where people can trade things they no longer need," the CEO began, his voice a mix of excitement and caution. "But that community only works if you trust the people on the other side of the screen. That’s where AI comes in – not as a magic wand, but as a practical tool that helps us keep the promise of safety."

He went on to describe three concrete ways AI is being used right now. First, a suite of machine‑learning models scans every new listing the moment it’s uploaded. The system looks for patterns that typically signal a counterfeit product, a scammer’s copy‑and‑paste description, or even a photo that’s been lifted from a competitor’s site. If something smells off, the post is flagged for human review – a hybrid approach that the CEO says balances speed with judgment.

Second, Carousell has rolled out an AI‑driven chat‑assistant that pops up during negotiations. It can suggest polite phrasing, remind users of the platform’s payment protections, and even warn them if a buyer’s language starts to look suspicious. "It’s like having a helpful friend whispering in your ear, nudging you toward safer choices," he chuckled.

Third, the company is experimenting with predictive fraud detection. By analyzing past transaction data – how often a seller changes their price at the last minute, the speed at which a buyer pays, the geographic distance between parties – the algorithm can assign a risk score to each deal. High‑risk transactions trigger an extra verification step, such as a request for a photo ID or a short video call.

All of this, the CEO admits, is still a work in progress. "Artificial intelligence isn’t perfect. It makes mistakes, it learns the wrong thing if we’re not careful. That’s why we keep a human layer in the loop, and why we listen closely to user feedback." He recounted a recent incident where the AI mistakenly flagged a legitimate vintage camera as a counterfeit. The seller received a polite email explaining the error, and the issue was resolved within a few hours. "Those little hiccups are inevitable, but they teach us to refine the models," he said.

Beyond the technical side, the interview touched on the emotional aspect of trust. The CEO believes that when users feel protected, they’re more likely to stay, to invite friends, and to post higher‑quality listings. He shared a story of a mother in Jakarta who sold a gently‑used stroller after the AI‑assisted system confirmed the buyer’s credibility. "She told us she felt at ease because the platform walked her through every step," he recalled, smiling.

Looking ahead, Carousell plans to expand AI capabilities to include visual similarity detection – basically, teaching the system to recognize when two photos are of the same item, even if they’re taken from different angles. This could dramatically cut down on duplicate listings and reduce the noise that often frustrates shoppers.

Still, the CEO emphasized that technology alone can’t solve every problem. Community guidelines, clear policies, and swift customer support remain essential pillars. "AI is a force multiplier, not a replacement for good people," he warned.

In short, Carousell’s strategy is a blend of sophisticated algorithms and human empathy, aimed at making the marketplace feel a little less like a gamble and a bit more like a trusted neighborhood garage sale. As the CEO put it, "If we can get a user to close a deal and walk away feeling good about it, that’s the real success – and AI is helping us get there, one transaction at a time."