Forum Discussion
Solved
Starting with Syntex - Hierachy Question
Hey everyone,
i am currently trying to establish a contract management system in Sharepoint Syntex and I am a bit lost. We implemented Office 365 and still struggle with setting up the Sharepoint how we would like it.
So I am the one person that is going to try to get "his part" done and then show how it would/could work.
Now I am trying to use the Syntex to get the contracts and other documents somewhat sorted which leaves me at a dead end.
The creating modules part is done and i think i got the kinda hang of it. What however is really confusing is the "document library" part in the Content Center.
1) Do i need to upload the files all in the content management area in order to use the models?
e.g. I have a Teams User "Finance" who has a sharepoint library called "Finance" and a subfolder called "Invoices". I therefore want to give them a list of all documents that are filtered with the model (e.g. Invoices). So Syntex creates a library called "invoices" but can i also upload the files in the original subfolder from finance?
Do i have the files twice then or does Syntex just extract the infos and "links" the files in the original folder?
2) List:
When applying the models/syntex to the library can i create a "listing" that is only visible for the department? E.g. Finance only for finance, Purchase only for purchase?
Thanks in advance
- Document understanding models that you create in the content center do not need to be applied to the document library in the content center but are applied (as you mentioned above) to other document libraries. Let me see if I can summarize:
You create the model for "Contracts" in the content center. Once your model is tested and ready to publish, you publish the model to the Finance Document Library for your finance user. Document Understanding models are deployed to a document library (not a folder although that may be a nice feature request for the future). The finance user will upload content to the folder and those documents will be evaluated by Syntex and the Contracts model applied if it matches the classification tests for the model and then the metadata extracted.
If you have more than one model applied to a library, all of the models are evaluated in parallel and if multiple models match the content the one with the highest classification score will be used.
If you want to apply a model to multiple document libraries, that is also possible for Document Understanding Models. Just publish the model to multiple libraries and you are good to go.
For your second question about a "listing or view", you can always use the contentype to return the content that matches (e.g., one view for Purchase that returns PurchseContracts and another for Finance that returns FinanceContracts) if I understand your meaning. These are not permissioned views though but more of a filter. If you want to have permissions, have one folder for each ContentType and either have people upload to the right folder or us a Flow to move the document to the right folder with metadata after it is classified and extracted by Syntex.
Hope this helps.
- Document understanding models that you create in the content center do not need to be applied to the document library in the content center but are applied (as you mentioned above) to other document libraries. Let me see if I can summarize:
You create the model for "Contracts" in the content center. Once your model is tested and ready to publish, you publish the model to the Finance Document Library for your finance user. Document Understanding models are deployed to a document library (not a folder although that may be a nice feature request for the future). The finance user will upload content to the folder and those documents will be evaluated by Syntex and the Contracts model applied if it matches the classification tests for the model and then the metadata extracted.
If you have more than one model applied to a library, all of the models are evaluated in parallel and if multiple models match the content the one with the highest classification score will be used.
If you want to apply a model to multiple document libraries, that is also possible for Document Understanding Models. Just publish the model to multiple libraries and you are good to go.
For your second question about a "listing or view", you can always use the contentype to return the content that matches (e.g., one view for Purchase that returns PurchseContracts and another for Finance that returns FinanceContracts) if I understand your meaning. These are not permissioned views though but more of a filter. If you want to have permissions, have one folder for each ContentType and either have people upload to the right folder or us a Flow to move the document to the right folder with metadata after it is classified and extracted by Syntex.
Hope this helps.First of all thank you very much for your answer and time to reply to this.
I just managed to get the second content center and start from scratch (it is a testing enviorment).
I just created two different models - one for contracts one for invoices. From what i gather i would now share the contract model with our legal department and they in return apply it to their library.
Now where does that information show up in the end? I let the model run trough and i saw that my sharepoint now has a column "classification" but that remains empty despite the same contract in the training gets labeled/classified correctly.
Also; Is there a central "hub" for the results to be returned to?
At the end of the day the main goal is to have "somewhere" all the information is available at a glance ( a list) and not by me going trough each and single library, as i for example want to put reminders in place to warn me that a contract runs out.
This centralized view works for the "invoices" which i uploaded to Syntex Content Center but not for the ones i have uploaded somewhere in the library. It just adds the infos there.
- The central content center contains information about the models and where they are used but not the metadata and content for the documents processed. The best way to do that is by using a search webpart using the various content types that match your models.
