IMPORTANT INFORMATION:

Basware is rolling out a new feedback method based on changes made during SmartPDF's self-validation phase and in AP. Please contact Basware's support team or your Customer Success Manager to confirm whether your setup is eligible for the new Touchless Feedback Loop.

Additionally, AP customers can check the new Insights - Invoice data quality report (Essential/Pro).

The following article provides background on the data capture service quality feedback process, the improvements applied, and the specific sub-processes for each analyzed root cause and improvement step. 

What is the data capture feedback process?

The data capture quality feedback process is intended to help customers and Basware improve the results of the Scan & Capture Service, SmartPDF, and PDF e-invoice services. Data capture is the common name for these services.

The customer collects feedback regularly and submits it in a feedback form. This form contains invoice information, the incorrectly captured fields, and expected field values. It is then sent to Basware for analysis.

Two different ways to report feedback:

• AP Automation customers use the built-in feedback reporting function by adding error codes to the documents.
  • The AP user's logged feedback can be exported from Insights.
• Some of the AP Automation customers are eligible for Touchless Feedback Loop, where the changes made to the documents in AP are sent to SmartPDF as feedback to be utilized automatically to improve capture quality.
  • Quality can be reviewed at the supplier and field level with Invoice data quality report - Basware Customer Support
• Customers without a Basware invoice processing system can report feedback by filing a manual feedback form.

Based on the customer's reported fields, the feedback analysis team identifies where improvement can be made. The reported errors will be analyzed thoroughly to find the underlying root cause and possible actions to achieve long-term quality improvement.

The analysis results are a set of key focus areas for improvement.

• The Scan & Capture process consists of numerous steps, from sending the paper invoice to appearing in the invoice processing system.
• In SmartPDF, the business document can be processed through various data-capturing engines.

The improvement action can be identified anywhere in this process.

Each analyzed feedback form contains an owner for each identified improvement action. In addition, there will be a unique ServiceNow identifier to track progress for errors where a feedback follow-up is required. The follow-up ticket (incident) will be opened under the same ServiceNow user who has reported the originating feedback. Reported errors where a follow-up is opened are called non-capture errors outside the data capture. A follow-up will not be opened if there are fewer than five documents, but the analyzed feedback report links to the specific Knowledge Base article.

Feedback process areas

The data capture service can be divided into two methods: automated capturing and human validation.

The following are the key processes:

• Field category based analysis vs. Individual invoice analysis:
  • Feedback processing per processing engine:
    • AI - All reported documents are used for AI training.
    • Template - Each reported document is categorized by sender ID, and the top 30 senders are fixed based on the highest sender document count in the feedback report.
    • Human validation by partner - Instead of answering each invoice feedback row separately in the human validated flow, we will focus on identifying patterns of recurring issues in field categories. We will focus solely on the top three field categories with the highest volume of impacted documents in the single feedback report, aiming to resolve the key issues that bring the most value.
• Proactive support ticket creation and tracking
  • A subset of improvement actions will be tracked in separate ServiceNow tickets (incidents), which makes the process more proactive, transparent, and simple.
    • ServiceNow has two ticket types: cases and incidents. Basware opens incidents, while customers and externals open cases.
      

Overall quality feedback process summary

This section provides a more detailed overview.

Here are definitions of the key aspects of the report:

• Field Category - This represents the document field that has the majority of quality issues associated (e.g., Invoice #)
• Ratio - The overall percentage of quality issues related to this field
• Error Category - This attribute provides a list of potential error categories for which the specific field errors are associated
• Finding - This is the fundamental reason why the issue is occurring
• Action Required - This represents the action required to resolve the identified root cause
• Responsible Party - This represents the party responsible for implementing or triggering the activity highlighted in the action required section 
• Support Ticket # - This section may provide individual support tickets; this depends on the action required
• More Information - This section contains a link to the Basware Knowledge Base. The article will provide a comprehensive summary of work instructions associated with the Finding and Action Required. We encourage users to check in periodically for updated content.
  
• Example - This section contains an example of the Finding to illustrate the reported issue further 

 For comprehensive details about the different data capture services findings, please review the following articles:

• Change Request required
• Configuration error
• Human validation mistake
• Service update required
• Master data error
• Scanning Center Error
• Data missing from image
• Image poor quality
• Example too old
• Expected value released
• Captured according to DCG/CMOS/SVG 
• Insufficient details provided
• Duplicate invoice
• Uncategorized error
• Wrong business unit selected
• Text extraction

 

Useful links:

• Export your feedback report from AP Automation via Insights
• Invoice data quality report - Basware Customer Support