What is Zero Trust? - Let's start with the definition
Zero Trust is a transition from implicit trust—assuming that everything inside a corporate network is safe—to the model that assumes that the corporate network has ALREADY been breached. Since you can not assume any safety from any network point (either externally or internally, you must explicitly verify the security status of identity, endpoint, network, and other resources based on all available signals and data. By using real-time policy enforcement to achieve the LEAST privileged access and minimize risks.
Wow! - That's quite a lot - but if you break it down into smaller, easier to understand bites, the idea becomes more manageable.
First - You must assume that the hacker is already in your network. Yes, that is a very scary first step, but it is the most critical. It is not a question of "if" anymore, it is now a question of "when" the breach will happen. If you start with that ending in mind, you can be better prepared.
Second - By utilizing a combination of Artificial Intelligence for pattern recognition, with checkpoints that are monitored by veteran security experts, ALL data is examined, regardless of the source or destination.
Thirdly - Even if the data is verified, the source is checked, the destination is checked, there is only enough access granted to do the absolute minimum that is required to perform the task. Think of the door or window only being opened just far enough to let the task complete, then immediately being closed again.