The app analyses the document text and matches rules based on keywords.


Keywords


Keywords need to be unique for a contact and should not appear in the document text for other contacts. A good starting point for a keyword is the contact name like "Joe's Flowers".  An example of a keyword that is not unique would be e.g. "American Express". The term "American Express" can be located on an invoice from the company American Express. It also can be located on invoices from other contacts e.g. in the sentence "We accept American Express".


The keyword does not necessarily need to be the contact name. Sometimes it is not possible to match the contact name e.g. when the contact name is part of a logo image.


Good candidates for unique keywords are:

  • Contact Name
  • Phone Number
  • Email Address
  • Business Number (e.g. ABN in Australia)


The keyword matching works best if you enter the keywords exactly as shown in the document text.


Keyword Matching


By default all keywords need to be found in the document text to match a rule. For example, if you enter the keywords Maria's Dry Cleaning, the words "Maria's", "Dry" and "Cleaning" all need to appear in the document text. The words do not need to be located next to each other as long as they are all present in the document text.



Phrase Matching


If you want to match the exact phrase Maria's Dry Cleaning, you need to put the phrase into quotes like this: 


"Maria's Dry Cleaning"


Putting the phrase into quotes leads to only documents being matched that contain the exact term "Maria's Dry Cleaning" with the words being located next to each other.



Multiple Phrases


You can also enter multiples phrases like this:


"Maria's Dry Cleaning" "Credit Note"


This will match documents that contain the phrases "Maria's Dry Cleaning" and "Credit Note".



Match Type


The rule section allows to select a match type from the dropdown box.


Available match types are:

  • Match All Terms
  • Match Any Term


Match All Terms: This is the default match type and means all terms need to be present in a document to be matched e.g. "Maria's Dry Cleaning" "Credit Note" means the document text needs to contain "Maria's Dry Cleaning" and "Credit Note" to be matched.


Match Any Term: This match type will match a document if the document contains any one term. For example the phrases "Mario Brothers" "Mario Bros" will lead to documents being matched if they contain the phrase "Mario Brothers" OR "Mario Bros". It is not necessary that the document contains all phrases as long as the document contains as least one of the terms.



Verify that a rule has been applied


Once an invoice is uploaded into the inbox you can check if a rule has been applied by checking the automation tag in the inbox. The automation tag has an automation symbol with the name of the rule next to it. The screenshot below shows an example where a rule named "Amazon" has been applied to a document. If you click on the automation tag you will be taken directly to the corresponding rule.




What to do if a rule is not applied


If a rule is not applied (and there is no automation tag/symbol in the inbox) please check the if the document actually contains the keywords exactly as they appear in the keywords section of the rule. The keywords need to be present in the document text. The app cannot extract text from a logo image.