Intelligent Virtual Assistant: Frequently asked questions (FAQ)
Intelligent Virtual Assistant (IVA) is a feature that uses advanced technologies like machine learning to extract invoice information from documents in your Inbox. The information we'll attempt to extract is:
- Invoice #
- Invoice date
- Due date
- Vendor name
- Payment terms
The information will be pre-filled on the review bill screen and the bill will only be created after you review and accept the information.
- How does Intelligent Virtual Assistant work?
- What is Click and Capture?
- What else should I know about this feature?
- Why did Intelligent Virtual Assistant not read all fields of the bill?
- Is there a limit to how many pages a multi-page document can have for this feature?
How does Intelligent Virtual Assistant work?
As PDF, png, jpg, or jpeg documents of up to 50 pages come into your Inbox, Intelligent Virtual Assistant (IVA) will begin detecting details on the document within seconds.
Documents with 1-2, or 21-50 pages:
On documents with 1-2, or 21-50 pages, IVA will attempt to detect details from only the first page.
Review and save
After IVA detects info on a document, select Review and save to fill in any missing info and save the bill.
If you need to create a new vendor for the bill, select Add new vendor. IVA will try to prepopulate the following new vendor info from the document:
- Vendor name
- Vendor email address
- Vendor full address
You can also use an existing vendor by selecting it from the Vendor Name dropdown.
Once you have reviewed or edited the populated fields and entered any information that is missing, select Save to create the bill.
Documents with 3-20 pages:
On 3-20 page documents, IVA will attempt to detect if there are multiple bills in the single document by comparing vendor name, invoice number and due dates on the pages of the document.
If IVA detects multiple bills, you'll see a notification on the document in the list that says Multiple Bills, as well as a banner on the preview pane. Note: You'll only see multiple detected bills in List View of the Inbox
Below the main document you'll see the suggested separated bills listed. If a suggested bill is 1-2 pages, you can edit and save.
On suggested bills over 2 pages, select Enter bill to double check
If the detected bill separation is incorrect, select Undo Separation in the document preview. The listed separate bills will be removed, and the main document will remain in the Inbox for you to manually create bills.
What is Click and Capture?
For those documents or fields we aren't able to process or copy to the bill via IVA, we have Click and Capture, a feature enabling clickable copy and paste from the document image to the Bill Summary or Expense Details windows to easily complete the bill's basic details from the document.
Simply select text to copy from the bill's document image, then select the field in the Bill Summary or Expense Details window where you'd like to paste the text.
Ways to select text:
- Move your cursor around a block of text;
- Hover near a single word until the single word is highlighted to select and copy a single word
- Hover above a sentence until the full sentence is highlighted to select and copy a full sentence
- Hover above a paragraph until the full paragraph is highlighted to select and copy a full paragraph
- To manually copy multiple text regions at once, hold down the Command key (for Mac) or the Ctrl key (for PC), then paste in the appropriate field within the Bill Summary window
- To clear copied text without pasting, right-click or hit escape
Click and Capture also works when quick adding a vendor from the Bill Summary window. You can also click to capture an address from the document and paste into the vendor record for the vendor's address.
- This feature won't be available for screen sizes less than 768px width.
- Click and Capture does work on an undocked document
- Click and Capture won't work if a document is rotated
What else should I know about this feature?
Machine learning gets better over time. If we aren't confident with the accuracy of a document, we won't show you any results. Hence, you'll see some documents processed and some that aren't. For documents that aren't processed by IVA, they'll need to be entered manually.
Why did Intelligent Virtual Assistant not read all fields of the bill?
Machine learning algorithms will try to predict all required fields for creating the bill. IVA will only populate fields that have a high confidence rating.
Intelligent Virtual Assistant currently works on PDF, png, jpg, and jpeg type documents.
Intelligent Virtual Assistant doesn't currently recognize foreign currency, the amount field will be blank for foreign currency documents processed via IVA.
Is there a limit to how many pages a multi-page document can have for this feature?
Yes, Intelligent Virtual Assistant will trigger on a multi-page document up to 50 pages.