The AI craze that has been going on for the past few months is not going to stop anytime soon. And we're not talking about the latest Terminator movie. We're talking about the rollout of Parseur's new AI parsing engine in beta!
We are very happy to announce that a new AI parsing engine now powers Parseur.. This engine is able to parse your documents without the need for a template. It works simply by listing and naming the fields you want to extract from your documents.
We hope this engine will drastically reduce the time spent on creating and maintaining templates, saving you even more time and effort, as well as opening up new data extraction use cases that were not possible before.
Parseur AI engine: what you need to know
The new AI parsing engine is available in beta for all new and existing users. It is available for all plans, including the free plan. It is also available for all document types, including emails, digital and scanned PDFs, Word documents, etc. It can understand and extract data from documents in most languages.
The new AI engine works alongside the existing Text and OCR template-based engines, meaning that your existing templates will continue to work as before. You can still create new templates as before. In case a document cannot be parsed by any of your existing templates, once you have activated the AI engine in your mailbox, the AI engine will kick in and try to parse the document.
How do I activate the AI parsing engine?
There are several ways to activate the new engine:
- During the creation of a new mailbox, you will now have a toggle to activate the engine
- For existing mailboxes, you can activate the engine in the mailbox settings
- Finally, you can activate the engine in all of your existing mailboxes at once by going to your account settings. Please make sure to read and understand the existing limitations below before doing so.
How do I use the new AI engine?
The AI engine uses the field names from your mailboxes to look up the related data in your documents. So if your fields have a semantic name related to the data they're capturing, you shouldn't have anything else to do after activating the engine in the mailbox.
For example, if you have a field called InvoiceNumber, the AI engine will look for the invoice number in your documents and extract it.
To do so, there is a new Fields tab on the document view page where you can add and edit the field names:
If you have generic field names or names not related to the extracted content, you may not get the best results, and we recommend you rename your fields to something more meaningful. Give it a try, and let us know if you face any issues!
Two new predefined mailboxes: resumes and travel bookings
To celebrate the launch of the new AI engine, we have created two new predefined mailboxes for use cases that were complex to handle with templates: resumes and travel bookings.
- The Resumes/CVs predefined mailbox is able to extract the following fields from resumes or curriculum vitae: candidate name and contact details, education, work experience, skills, etc.
- The Travel Bookings predefined mailbox is able to extract the following fields from travel bookings: booking reference, trip segments, travel dates, passenger names, etc.
You will find those new options when creating a new mailbox.
Current limitations and gotchas
The new AI engine is still in beta, and we are working hard to improve it. Here are the current limitations and gotchas:
- Page count limitation: For now, the AI is capable of extracting data from up to about 10 pages of any document. The exact number of pages can be slightly more or less, depending on the text density of your pages. In any case, Parseur will not charge you more than 10 credits per document when using AI.
- No support for "Reprocess All": For performance reasons, we currently have to limit using the "Reprocess All" button to non-AI mailboxes only.
We will be trying our best not to change our pricing at this point, but we may have to do so in the future if we find that the new AI engine is too costly to run. We will keep you posted if that happens.