This information is only relevant to users working with parsing of emails or integrations to portals connected to Globus. If you create orders manually feel free to skip this section.
When working with orders from parsing (emails) or portals (integrations) data fetched by either the machine learning model or from the portal could have different formats, naming, and other structures than the data imported from your ATS system into Globus. Globus will attempt to find the matching customer, department, location, contact and role from the emails or integrations. In addition, it will attempt to extract due dates, shedules and requirements. A color coding logic is utilized to show missing information or uncertain information. The color coding is visible in the in the orders list. The colors show the certainty of the values parsed or extracted.
Green indicates high confidence score (Im very sure this is the correct value) and yellow low confidence score (Im uncertain if I found the correct value or I didn't find a matching value). By clicking the "AI" icon on the right side you will also see a written explanation of each value and also a preview of the email. The information is also available in order edit mode.
All orders with missing or uncertain information (yellow color) will open in edit mode for the user to confirm the details of the order.
As part of the onboarding of Globus, an important part is training the system to set the correct matching customer, department, etc automatically. One example could be roles, the customer could request an "RN" in the email while the matching roles or attribute used in your ATS/Globus could be "Registered Nurse". Globus does not know from the beginning what your matching roles are and as a user, you need to teach the system this. This is done by selecting the correct role in edit and confirming by saving the order. Then a feedback loop will be updated, meaning that the next time an order with role "RN" is received, "Registered Nurse" will be automatically set as the role on the order. The same logic is used for identifying customer and department. By selecting the correct customer and department, the system will learn and improve the identification.
Contact person is also used for identifying the correct customer and department. If the contact person is unique (only set on one customer and/or department), it will be used to set the customer and/or department.
By clicking an order in the list with missing information or uncertain information (yellow), the order will open in Edit mode. Edit mode will show you the email on the left hand side and all the order fields available on the right. A "*" next to a field shows that it is a mandatory field and it is not possible to continue working on the order before these are confirmed.
There is also an additional tab "AI processing" available at the top, by clicking it, you can see what data was extracted by the machine learning model and what values it selected from the data in Globus. Choose the correct data on the right-hand side and confirm it by clicking "Save order". By repeating this action, the system will learn and increase the likelihood of selecting the correct data.
