Companies Must Cleanse Data to Maximise AI Functionality

The adoption of artificial intelligence (AI) among companies is increasingly widespread, from chatbots to business process automation. However, many organisations are still mistaken in building the foundation for AI, particularly in data management. This view was expressed by Snowflake’s Regional Vice President and Managing Director for Southeast Asia, Satchit Joglekar, in an exclusive interview with Kompas.com at The St. Regis Jakarta in South Jakarta on Thursday (17/4/2026). Snowflake is a cloud-based data platform that helps companies manage, integrate, and analyse data to prepare it for analytics and AI needs. “AI is useless without data. Data governance is the most important thing. If the data is already organised and correct, only then can companies focus on AI,” said Satchit. He added that many companies today are too focused on using the latest AI models, such as generative AI (Gen AI) or chatbots, without first fixing their internal data. In fact, AI works by relying on data. If the data is not organised, not integrated, or scattered across various systems, the results provided by AI have the potential to be inaccurate. “In some cases, AI can even produce information that is not relevant to the company’s business conditions,” said Satchit. In AI implementation, companies often face the phenomenon of data silos, a condition where data is stored in various separate systems or storages. In addition to being scattered, these storage systems are often isolated, making data difficult to access or integrate with other systems. “This kind of condition becomes a major challenge for companies that want to adopt AI. Because AI requires data that is easily accessible and integrated in one place,” said Satchit. Not only that, data within companies is often not uniform. For example, certain terms or definitions can differ between divisions, even though they refer to the same thing.