The dreaded "New Template Needed" status means that Parseur could not find any matching template to extract data from the document or that the AI engine failed to extract data.
​

Before we see how to fix parsing failures, we'll look at some best practices to ensure smooth and reliable parsing.

When you use the AI engine to extract data and it fails, the document is marked with the "New Template Needed (AI)" status error.

There can be several reasons why the AI failed to parse your documents. Click on the 🔍 Logs button to see the log and get more details on the error.

The two most common errors are:

If you're parsing emails and text documents, they can sometimes be finicky to work with. Here are some best practices to reduce parsing failures when using text templates:

Manually forwarding emails to your Parseur email address can reduce parsing reliability because email clients tend to change email contents during forwards.

Parseur uses what is immediately before and after each field to locate them in the document. If you have to change the text surrounding a field, Parseur may not be able to locate that field and will trigger a "New Template Needed" error.

For example, if your template has Date: 1/2/2020 12:34 and you create a field to capture just 1/2/2020 as the date, you may have a problem as the timestamp 12:34 that follows the date is likely to be different when the next document comes in. In this example, you could either create another field to capture the timestamp or create a field to capture the full date and time and use Parseur field formats to format the date as you need.

Check out our Understand how Parseur works article to learn more about this and how to fix it.

Parseur requires that all fields match for a TEXT template to be used.

If a document wasn't parsed because one field wasn't present, create a new template based on this document. When you have more than one template, Parseur will automatically pick the best one every time.

Parsing PDFs is generally more straightforward than parsing emails, thanks to our new OCR parsing engine. Here are a few things to keep in mind:

If some of your fields move vertically or horizontally from one document to the next, use labels to help Parseur find them.

Read more about using labels to dynamically position fields (Dynamic OCR).

Fields are mandatory by default. You can set a field as optional by toggling the "Field presence is required" button off.

If you have implemented all the template best practices above and still encounter issues, here are the various ways you can fix failed document processing.

This is often the easiest and best solution. If your document layout is different from the previous one, create a new template to tell Parseur where to find the fields in this document.

To create a new template, either:

When you have more than one template, Parseur will automatically pick the best one every time a new document comes in.

If you are having trouble with an existing template not matching the document you want to parse, you can use the template debugger to identify the specific parts of the document that Parseur failed to match.

To access the template debugger, either:

In the debugger, you will see the source content of the current template compared to the document:

You can hover parts of the text with your mouse and see the corresponding parts matched in the document.

You need to have a 100% green for a template to match a document and extract its data. which means there should be no red pieces. To fix a template error, you can either add a new template or edit an existing one. If you are using text templates, you can read more about understanding how Parseur text engine works for additional information.

The AI engine is our latest and best data-parsing engine. It may be able to parse data where our templates cannot. Check out the pros and cons of AI vs templates article.