You do this inside your command handler. You save your domain objects as normal. Even if you are not using event sourcing as a means of persisting your domain entities, you will still need to fire events that your query service will subscribe to. The event handlers on the read side will then update de-normalised tables tailor made for the UI screens. So basically you have two sets of data access code: one for the domain, one for the query service (read side). Its less work if you make use of event sourcing to persist your domain entities...