When should I use File Content Extractor over an AI solution?
- legigs
- May 22, 2025
- 2 min read
QWERTY Software Solutions has released the File Content Extractor ("FCE") to the world, to fulfil a specific niche of needing to find and extract text from within large numbers of documents. A common question heard is, "can't we just do this with AI?"
AI text extraction is heavily focused on the training of large language models, and particularly good when you have a small amount of text (for example, a long paragraph) which you can copy / paste into a large-language-model and ask it do something with it (for example, "summarise this down to one sentence"). This is not the use-case for FCE.
Imagine you have a file server with half a million files on it, and what you want may be very esoteric (e.g. a legal firm locating evidence for a specific case) or perhaps something simpler, but you need to process the results in a particular way (e.g. a marketing team extracting email addresses to put into their CRM).
Here are just some of the challenges with the AI example provided:
Data security - those files may contain sensitive information you do not want to upload to an AI tool, hosted by an external vendor
Data transfer - there are many gigabytes of files which is difficult and time consuming to transfer individually
Training - in the example, a legal firm knows what they're looking for whereas AI may need a model trained to identify it independently
Workflow - the AI solution may do a good job of identifying the text but building operational and governance processes around it may be difficult without development
Not all text may be structured - LLMs require text as input, so particular documents like scanned images as PDF's may not be supported
To counteract these challenges, FCE:
Runs locally in your environment as a desktop application. We are not a SaaS provider!
Leaves the data where it resides (the user and computer running the application only requires read access to the files)
No model training is required. Limited testing is recommended to refine your search criteria, but the ability for FCE to read the files does not change.
Basic workflow is supported by allowing users to pick when they run scans and have FCE automatically copy or move the files to another location based on the outcome. It can optionally generate reports for governance and traceability.
Optical Character Recognition (OCR) is supported for image files, PDF's and Word documents. It also searches attachments within email files.
As an introductory offer (until 31 December 2025), in addition to our standard practice of a free demonstration, we are doubling the license offer free of charge.
3-month license ($900 per computer + user combo) is upgraded to a 6-month license
6-month license, ($1,440 per computer + user combo) is upgrade to a 12-month license
12-month license ($2,304 per computer + user combo) is upgraded to a 24-month license (not usually available!)
