Optical Character Recognition – or OCR – is a valuable application of AI. In short, AI OCR allows users to quickly scan documents and highlight important information. Unlike traditional OCR systems, ones powered with AI are able to decipher context to accurately recreate a document, even if it’s handwritten.

Workflow:

  1. AI OCR process will convert both printed and handwritten documents into text files.
  2. Documents can be added from existing saved documents or a sheet-fed scanner.
  3. Documents are saved to a shared drive.
  4. Once saved, the document management software begins processing.
  5. Processing the files adds tags to start the AI analysis.
  6. The document will then be added to a queue for tagging, generation of a title based on content, and removal of secrets.
  7. A creation date will be generated based on the latest date printed on the document.
  8. A redacted version of the text will be created.
    1. A separate system removes confidential information.
  9. The document will then be tagged for manual review to ensure tags and redactions are appropriate.
  10. Once reviewed by a human, it will be searchable by users via permissions based tagging.

AI functionality:

  • Removal of sensitive information by interpreting the context of the document.
  • Pattern and number recognition: SSNs, phone numbers, keywords, etc.
    • The agent permanently covers or crops sensitive information.
    • Permanent removal prevents sensitive data from being leaked.
  • Users can ask the agent questions from documents based on search results.
    • Optionally, an external AI agent with more modern and advanced systems can be implemented.
  • Creating a custom branded interface.
  • Integration with the company’s existing SSO.

Hardware needs:

  • Can work with existing IT infrastructure.
  • Can set up required hardware on site.
  • Can set up a server and/or backup redundant system depending on the number of pages/documents.
  • Can work with an existing IT department to come up with a co-location.
    • Offsite backups to be done per compliance requirements.

Benefits of Local AI:

  • No information is transmitted to the internet. Titles, tags and secrets are stored locally.
  • Local AI search queries go directly to the custom AI agent and do not have to be saved.
  • Users can decide whether or not to use AI configurations for the internet for queries. This depends on the level of security and desired skills for the agent.
  • Local AI potentially uses less water and resources than other agents.
  • Less data entry makes for a more efficient workflow.
  • The agent is scalable.

Potential drawbacks: 

  • May be slower than desired based on hardware.
  • Higher cost of acquisition.

A+L Development has been working with AI since long before it was a buzzword. Reach out today to find out how we can use AI to make your business more efficient!

TAGS
Software