[Music] let's see a demo of document intelligence for CSM in Tokyo our business is struggling to handle a surge of inbound vendor invoices as a system admin I'm interested in helping to improve vendor response and streamline operations in filter Navigator type task Intel and select document intelligence this will take us to a list of task definitions select handle vendor invoices the task definition are the various configurations that enable us to hook into an AI model in this example we are setting it to autofill meaning to autofill the various fields on the case form we're enabling straight through processing against the target table of case under Keys these are the various configured Fields namely vendor name invoice number invoice date amount due and tracking number and you can see how they're associated with the various Target fields on the case as the target table under tasks here are the various sample documents that we have done some pre-training of the model let's go ahead and click on Sample dock one and click on show in Doc Intel on the first pass of this first document you can see that the model hasn't done such a good job but we can help it out in fact the model has identified all the fields as being customer order number let's go ahead and correct the model by suggesting that the vendor name should be boxio and that the invoice number should be 8675-3099 and that the invoice date should be July and that the invoice amount should be 540 and that the tracking number is actually not present uh in the document let's go ahead and submit that for sample dock number two we click on show in dock Intel and on this second document the model is starting to get smarter already you'll notice that it was able to identify a vandelay industry and it was able to identify the invoice number and the invoice date however it incorrectly identified the amount due so let's go ahead and help it out a little bit and we'll say that the amount due is 107.22 and once again tracking number is not present or is missing in the document let's go ahead and submit that so we will go ahead and click on Sample dock 3 and we will click on show in document Intel and you'll notice that the model uh got smarter it's identifying that Freight boy is the vendor the invoice amount the invoice date the amount due of being 12.50 and once again the tracking number is not present on the document the other thing that's noteworthy is that this document has a slightly different format and it's actually skewed a little bit to the right however the model was still able using this OCR capability to protect to pick up those data fields so we'll go ahead and we'll submit that and now we have our model pre-trained the last thing to consider is the different conditions upon which we want to have this model fire so now that we've trained the model against those three sample documents you can also specify what we call integration setup and under integration setup this is where you can Define The Logical condition that will determine when the model fires and in this example what we're doing that the short description contains the word invoice and that will determine whether this model fires so now that we have our model set up we have defined the autofill the target table the various Fields we've traded against three documents and we set up the integration steps our next task is to actually send an email from a company called Vortex Doors and they are submitting their invoice to Acme Corporation and retrieve the email within configurable workspace you'll notice that the email has manifested itself as case number 1725 we will click on that immediately on the case form we can see the fields that the model was able to predict so it was predicting the account name as boxer USA is predicting the vendor name the invoice the invoice number and the amount due let's review this in Doc Intel and you'll see that all four Fields were successfully identified so the vendor name the invoice date invoice number and perhaps most importantly the invoice amount of 107.22 so we have seen the document intelligence is able to deliver tremendous value to organizations it helps eliminate OCR and RPA solutions that are difficult to administer and maintain it also helps to digitize manual inspection and eliminate data entry and this will streamline operational efficiency and enable support teams to respond to their end customers quicker thank you very much