You can convert invoices into structured data in Cloud Data Fusion
using the Invoice Parser plugin, which is powered by Document AI. The
structured data gets stored in BigQuery.

## Before you begin

To parse invoices, you need a Cloud Data Fusion instance running in
version 6.4.1 or later. For more information, see
[Upgrading Cloud Data Fusion instances](https://docs.cloud.google.com/data-fusion/docs/how-to/upgrading#upgrade-instances).

## Create a processor

1. In the Google Cloud console, go to the Document AI **Processors**
   page.

   [Go to Processors](https://console.cloud.google.com/ai/document-ai/processors)
2. [Create a processor](https://docs.cloud.google.com/document-ai/docs/create-processor). Select **Invoice
   parser** as the type of processor.

   ![Select Invoice Parser as type of processor](https://docs.cloud.google.com/static/data-fusion/docs/images/how-to/parse-invoices/invoice-parser-tile.png)

   > [!NOTE]
   > **Note:** When you create the processor, copy the processor ID for your plugin configurations.

## Configure the invoice parser plugin

1. In the Google Cloud console, go to the Cloud Data Fusion **Instances**
   page.

   [Go to Instances](https://console.cloud.google.com/data-fusion/locations/-/instances)
2. Ensure that the desired instance has been upgraded to version 6.4.1 or
   later. For earlier versions, [upgrade the instance](https://docs.cloud.google.com/data-fusion/docs/how-to/upgrading).

3. Click **View instance**.
   The Cloud Data Fusion UI opens.

4. Click **Hub**.

5. Click **GCP** , and then deploy **GCP Plugins**.

6. Click **DocAI** , and then deploy the **Doc AI Plugins**.

7. Click the **Invoice Parser Quickstart** \> **Create**.

8. Customize your pipeline by entering the Invoice Parser processor ID,
   Cloud Storage bucket path, and BigQuery table details.

9. Deploy and run the pipeline.

   ![Example pipeline with Invoice Parser plugin](https://docs.cloud.google.com/static/data-fusion/docs/images/how-to/parse-invoices/invoice-parser-pipeline-example.png)

Parsed invoices are stored in the output table in BigQuery.
Metadata from the invoices is stored in the `Metadata` table and includes
parsing status, Cloud Storage path, and upload timestamp of the raw
invoice. Records in the output and metadata tables can be joined with the
`invoice_uuid` key.