Maybe Dilbert's Employer Will Try To Get This Contract

Below are extracts from a request by Social Security for information from potential contractors. The agency is not ready to seek bids, but is seeking information only.

Let me make it clear that I regard this request as a complete fantasy that reveals the naivete of whoever came up with this idea and whoever approved this request for information. What is requested is far beyond the capacity of any computer program -- but some contractor will certainly be willing to promise to produce a program that does this, for $100 million or so, but definitely without a money back guarantee. Go ahead with this nonsense and you will be wasting lots of money.

The Social Security Administration (SSA), Office of Systems (OS) and Office of Disability Adjudication and Review (ODAR) are seeking to identify qualified sources interested in providing and implementing a Commercial-Off-The-Shelf (COTS) Image Knowledge Base software product or a service that provides and/or utilizes the functionality of an Image Knowledge Base software product. The software or service may consist of multiple components, if necessary to meet the SSA's requirements. ...

ODAR employees in hearing offices classify documents into various categories and extract pertinent information from those documents. The employee must look at a document's identifying features, determine the type and title of the document and then enter the information into the electronic folder. Afterwards, the employee organizes the documents chronologically and removes duplicates. ...

The SSA invites all qualified sources to respond to this sources sought announcement.ALL REQUIREMENTS ARE MANDATORY.

REQUIREMENTS:

• At a minimum, possess the capability to process a national workload from 145 hearing offices and remote sites nationwide that process approximately 550,000 cases per year that consist of, on an average, 300 document images per case. Possess the capability to process these images within an acceptable period of time and with an acceptable level of accuracy. ...
• Possess the capability to use the knowledge base of information to classify the document images into categories and subcategories according to the established SSA Electronic Folder (EF) rules for classification.
• Possess the capability to separate document images from one another to avoid multiple document images from being inappropriately scanned, stored, and treated as a single document image. The document images may be any combination of multiple known forms and/or multiple free format text documents. This needs to include the ability to address those document images that already exist within SSA's unstructured data repository where there is already some document separation and classification information.
• Possess the capability to extract structured data from the document images such as Social Security Number (SSN), treatment dates, treatment source names, etc., and place the data in the appropriate fields specified by SSA for the Electronic Folder using SSA's application program interfaces (APIs)....
• Possess the capability to automatically organize the document images (after classifying them), chronologically by treatment date or other parameters established by SSA for the EF.
  
• Possess the capability to support SSA's viewing function (of the document images) by accessing the stored metadata in the EF, using SSA's API's, to organize the document images based on parameters established by SSA.
• Possess the capability to identify and flag duplicate document images.
• Possess the capability to identify and flag document images with poor quality.
• Possess the capability to identify and flag images with low probability of being classified correctly with the ability to configure and dynamically adjust the probability ratio.