Our founder’s first exposure to the importance and need of data protection, as well as its consequential implications to data science workflows was experienced during his research studies when data access became a bottle neck for progress and barely resolved during his “write up” year. As a consequence, another limitation was quickly realised when developed AI algorithms validated on open datasets did not generalise well across datasets/organisations. These two problems: (1) data access/protection and (2) AI algorithm generalisability became the seeds of the inception of DeepCognito.