AI effectiveness depends on data quality, warns Columbus EIM

Joakim Jannerfeldt, Vice President & Head of Columbus EIM, warns that the effectiveness of artificial intelligence is strictly limited by the quality of the information used to train it. Many companies are investing heavily in AI but risk reinforcing existing organizational problems by training models on unstructured or incorrect data.

When documents, data, and processes lack structure—often scattered across various business systems, emails, and local folders—it creates a risk of employees working from inconsistent information. This leads to duplicate work, longer lead times, and unreliable decision-making. For AI to provide relevant and useful answers, it must be based on information that is correct, coherent, and easily accessible.

However, Jannerfeldt suggests that the need for a strong information foundation should not prevent companies from starting their AI journey. Instead, implementing AI can help organizations identify where their processes need development and where information must be structured to maximize business value.