By AIM · AIM Media House
Bad data costs the global economy an estimated $3 trillion every year. That’s more than the GDP of France. And as companies rush to embed generative AI into customer service, product recommendations, forecasting, and pricing, the risk of making decisions on flawed data is rising fast.
Sifflet, a startup based in New York, has raised $18 million to address that problem. Its AI-native data observability platform is designed to ensure the reliability of data pipelines before they cause real damage. Founded in 2021 and launched in 2023, Sifflet has since tripled both its customer base and its revenue.
Its clients now include Penguin Random House, Euronext, and Saint-Gobain. Pete Williams, Chief Data Officer at Penguin Random House, says the company uses Sifflet “to ensure data reliability across the business, not just within engineering.” Sifflet’s model challenges how most organizations think about data quality.
Instead of treating it purely as an engineering problem, Sifflet says their making it a shared responsibility. It delivers tooling for both technical and non-technical teams, aiming to embed data reliability into decision-making processes across the enterprise.
No AI Without Good Data The company’s CEO and co-founder, Salma Bakouk , argues that businesses can’t unlock the value of AI until they first address the integrity of their data.
Along with co-founders Wissem and Wajdi Fathallah , she built Sifflet to provide both visibility and control, without requiring teams to rewrite their workflows.
The platform includes standard capabilities like anomaly detection and data lineage, but also introduces AI agents that assist with incident response and resolution. Sentinel monitors metadata and recommends what to watch. Sage traces the origins of problems using lineage and query history.
Forge suggests fixes based on past resolutions.
Continue on AIM Media House