Talking AI — Synthetic data: training AI when real data is missing, only if it mirrors yours

Synthetic data is generated by a model rather than collected from the real world: it is used to train or test an AI when real data is missing, sensitive or too scarce. It unlocks projects, but is only worth anything if it faithfully mirrors the company's real cases, otherwise the model learns a world that does not exist. Gartner predicts that by 2027, 60% of data and analytics leaders will face critical failures in managing this data, putting governance, model accuracy and compliance at risk. The real question to ask a vendor is not whether it uses synthetic data, but what share it represents, how they validate it against your real cases, and who is accountable for it. Before letting any synthetic data into a project, require that it be checked against a measured real sample and that someone from the business sign off on that check.

Source: Gartner, "Gartner Announces Top Data and Analytics Predictions," June 17, 2025.