To get started with Rossum, pretty soon you will want to configure your Rossum account according to your needs.
We call your whole Rossum account an organization, where multiple individual user accounts may have access. You will want to add a user account for every individual accessing your organization, whether to do the actual data capture or administer the account.
The documents processed within your Rossum organization are flowing in queues, transitioning through a variety of states. Each queue may be configured to capture different data, have different user accounts associated, and allows separate export of data. Queues may be logically grouped in workspaces for ease of orientation.
You will want to go beyond your single queue if you need to separate your documents into multiple streams. That can be because you need to capture different information from various document types, will have multiple teams working on documents, or just because of how your business process is set up.
Moving to production
Having a separate testing and production queue is a good idea. If you are confirming documents in a queue without validating the captured data properly, that may confuse the AI and reduce the automation rate until the AI "unlearns" its mistakes again, which may take quite a while. Since the AI learning is confined to specific queues, handing the account over from developers to the data capture team may be a good moment to cut the history clean for the AI's sake.
The particular structure of the captured data is described in a schema. Basic tweaks of the fields to capture can be done in the Field settings section - you can add and remove extra standard fields to capture everything you need and not waste time on fields that are not necessary. More advanced tweaks can be done in the Schema Editor - adding custom fields and sections, controlling required and default values, or tweaking automation levels.
The easiest way to update all of the above is within the Rossum app itself in the Settings section. Sometimes, it is useful to apply a lower-level access to the API directly, e.g. when scripting around some operations - you will want to use our command-line configuration tool elisctl.
Updated 4 days ago