How Klarna turns half a million reviews into a real-time analytics engine

We sat down with Klarna CMO David Sandström at OMR Festival to talk about why open consumer feedback has become core infrastructure — for marketing, product, and staying discoverable in the age of AI.

In an era of large language models (LLMs), Klarna and Trustpilot are collaborating to prove that authentic human experiences remain a brand's most valuable asset. 

Here’s how Klarna uses that sentiment data to stay visible, trusted, and discoverable in the age of AI.

To drive change, not reflect past performance 

Some businesses treat feedback platforms as a passive record of yesterday's transactions. Klarna looks at Trustpilot as a real-time analytics engine.

"I used to work in consumer research and a gazillion dollars per year was spent on gathering focus groups and sending out surveys. What we're doing now is building synthetic audiences where we ask the AI to mimic certain audiences based on input data,” explains Sandström. To gather this data, Klarna has built a tool called the Trustpilot Review Explorer, a chatbot that draws on their half a million reviews and creates synthetic audiences based on input data from Trustpilot, general research, and customer support interactions.

“So, we would say, hey, what would a 35-year-old woman in Hamburg feel about this? How would they look at this design?” Sandström then uses the tool to speak with simulated consumers modeled on real data. This gives Klarna real-time feedback that predicts how real people might react to a new app layout or user experience before a single line of code ever launches.

“The wrong way to think about Trustpilot, in my opinion, is as a rearview mirror," says Sandström. "I see it as an early indicator of something – we’re reaching a certain threshold of people reporting specific issues or things they appreciate… so, I try to use it as a forward-looking tool."

To get AI recommended

The way buyers find businesses has fundamentally changed. Traditional search engines are rapidly losing ground to conversational AI platforms like ChatGPT, Perplexity, and Claude. But with AI as the new gatekeeper of search, how does it choose which brands get recommended?

AI models build their definitions of quality on consumer experience, meaning an open feedback ecosystem provides the transparent data layer required to be a top-cited answer. In essence, they read what people say about a business on authoritative, open platforms.

"Reviews are a primary source for AI answers because algorithms only know what we feed them," Sandström notes. AI search identifies the best options based on honest human consensus, which is why keeping a transparent footprint of consumer feedback is so critical for modern discoverability.

Watch the masterclass

Ready to win the new rules of AI search?  Watch the full OMR Masterclass for the full discussion and actionable insights on trust, data, AI search, and future consumer trends.