I've recently seem event storming as a quite reasonable way to get understanding form a domain, in order to build a domain model. We understand the events in the domain, which for the domain experts ends up boiling down to just telling the business processes they are used to in their everyday life, then we find out the commands that might rise these events, and finally the aggregates.
By everything I've read it seems a very viable way to understand the bussines rules and business logic.
But there are many points that I still don't understand. The first would be: when should event storming be used in the course of the project?
My doubt seems to be mainly regarding event storming and requirements gathering. Should it be done just after requirements gathering, or it should be viewed also as a technique of requirements gathering?
In summary, in what stage of the project should we use the technique of event storming?