Document AI Platform: How Affinda helps mitigate risk associated with handling personally identifiable information (PII)

Document AI Platform: How Affinda helps mitigate risk associated with handling personally identifiable information (PII)

Share this post
  • All documents containing PII identified
  • High risk documents secured appropriately  
  • Reduced risk exposure
March 25, 2024
5 minutes

Summary: 

In the aftermath of major Australian data breaches, Affinda's customer needed a solution to tackle the challenge of handling sensitive data. Affinda has developed an AI document processing solution that scans all documents and document-like artefacts in a database, and automatically detects and classifies documents with personally identifiable information (PII). For each document, the solution returns a document classification, a PII risk score, and an explanation of the type of PII detected. This allowed our customer to set up the appropriate security restrictions around these high-risk document types.

An example of Affinda's AI document processing solution, classifying a newspaper article.
An example of Affinda's AI document processing solution, classifying a newspaper article.

An example of Affinda's AI document processing solution, classifying a resumé/CV.
An example of Affinda's AI document processing solution, classifying a resumé/CV.

An example of Affinda's AI document processing solution, classifying a utility bill.
An example of Affinda's AI document processing solution, classifying a utility bill.

An example of Affinda's AI document processing solution, classifying an Australian Passport.
An example of Affinda's AI document processing solution, classifying an Australian Passport.

Background:

For over two decades, Affinda’s customer has been providing services for thousands of businesses and individuals across Australia. They receive more than 100,000 documents annually, a significant portion of which are documents containing sensitive PII.  

Business Challenge:

With technology rapidly evolving, a major challenge facing many Australian companies is the handling of sensitive data. The Optus data breach of September 2022 was possibly the largest to have ever affected Australians, with hackers gaining access to personal information of over 11 million individuals. According to the ACCC’s Scamwatch service, the stolen data included: names, dates of birth, phone numbers, emails and postal addresses as well as some Driver Licence numbers, Medicare numbers and Passport details. In the same year, Medibank was hacked, exposing over 9.7 million Medibank customer’s personal details.

These large-scale data breaches have prompted Australian organisations to reconsider their approach to managing and safeguarding sensitive data. The primary issue for larger organisations is the accumulation of documents from multiple sources over the past few decades, all stored in various databases and data lakes. Unfortunately, these companies lack knowledge about the content and classification of these documents, including whether they contain PII, making it difficult for them to access or utilise this information effectively.  

Business Requirements:

Documents are received from customers when they sign up for a service.  

These documents can be sensitive in nature, and could include utility bills, resumes, Passports and Driver Licenses.  

Due to this, Affinda's customer required an AI document processing solution that:

  • Can automatically detect if a document has PII
  • Is scalable, fast and efficient and can handle millions of documents and other artefacts
  • Provides enough information to our customer’s team to determine next steps with the document, e.g. if the document contains PII, it must be stored securely and protected from unauthorised access

Solution:

Affinda’s AI experts worked with our customer’s team to develop an AI PII scanning solution that can automatically assess the risk of personally identifiable information contained within documents. Our model uses the latest advancements in AI to:

  • Automatically scan a submitted document to return a confidence level for the document type, from zero to 100, e.g. the model might return a confidence level of 99% that a document is a report, or 90% confidence level that the document is a Passport.  
  • Assign the document to one of 25+ different classes, including:
    • PII classes, such as a Passport, utility bill, resume, bank statement and Driver Licence  
    • Other classes, such as a report, photo, letter, chart or contract
  • Provides a clear explanation as to the type of PII found within the document

With the risk score, document category and type of PII automatically provided in a report once a document is submitted, our customer’s team can now make an informed decision about where the document should be stored within their system. Not only that, but our customer was able to:

  • Submit their backlog of unidentified documents and receive a report with the above information for each document
  • Send all new or modified documents that the customer receives through to Affinda monthly, allowing us to return a classification and confidence score for each document. This enables the customer to gain insight into any potentially risky documents within their system, helping them assess the overall risk level of their system and database of information over time

Impact:

  • All documents containing PII identified
  • High risk documents secured appropriately  
  • Reduced risk exposure

Ready to see how Affinda’s AI-powered document processing products can help your business? Chat with one of our experts today.

Related Case Studies

Resume Parser & Resume Redactor: Supporting ATS providers to enhance accuracy, performance & combat bias in recruitment

Learn how Affinda supports ATS providers to combat bias in recruitment while enhancing accuracy and performance with our AI Resume Parser & Redactor.

Document AI Platform: Felix streamlines insurance document processing and reduces manual data input by 76%

How Felix streamlines insurance document processing and reduces manual data input by 76% with Affinda's Document AI Platform.

Document AI Platform: Makesure automates identity document processing and reduces check response times by 99%

A case study on how Makesure automates their identity document processing and reduces check response times by 99% with Affinda.

Explore how you can process your documents with our powerful AI.

Get in touch with our team of experts and find the best solution for you. Contact us for a free consultation call.