Specifications include, but are not limited to: 1.1 Provide a tool that uses modern techniques of OCR and/or Artificial Intelligence, Machine Learning and combinations, to read and extract data from student transcripts. The process may also be capable of reading and extracting data from other document types (high school transcripts, test scores, etc). 1.2 Provide a tool that can read various file formats in which documents are received, including scanned image (tiff, jpg, png) as well as PDF. 1.2.1 Tool or process must be able to accept meta-data with transcript/documents uploaded for data extraction processing (e.g. University provides University student ID with transcript to be processed) and return meta-data (e.g. University provided student ID) as part of the data extract delivered to the University. 1.3 Provide a tool that can successfully process transcripts without requiring preconfiguration by customer. 1.4 Solution must be capable of processing transcripts from nearly any college, university, or high school in the United States with a high degree of accuracy. The word “nearly” is used to recognize that there are a few institutions that produce transcripts that do not adhere to standardized grading or course formats. University intends to initially focus on implementation for college transcripts and keep the option open to work with high school transcripts, but not part of the initial project. University processes approximately 17,000 college transcripts per year and approximately 35,000-40,000 high school transcripts per year. Not required, but if a solution is able to perform any course patterns, derivations, calculations or categorizations for high school transcripts that would be welcome. 1.5 Solution may leverage a shared library of algorithms so that the intelligence of transcripts trained from one customer becomes shared among all customers. Other methods of data extraction intelligent processing may also be used. Please describe. 1.5.1 Explain what tools or processes are provided to train the system when a new document type or transcript is required, and what change control mechanisms, if any, are provided for “teaching” the software. 1.6 Solution is preferred if it provides an intuitive administrator interface to view and review processed transcripts. If it does not, please describe how confidence scoring of extracted data and possible administrative or quality assurance review and/or editing of extracted data would occur. 1.6.1 User-friendly solution should allow administrators to begin using it with minimal training. Explain how administrator and processing user accounts would be created and maintained. 1.6.2 Administrators must be able to easily choose between multiple alternative data interpretations if the identified result is not optimal. 1.6.3 User interface should allow corrections and cleanups by end user before submitting the data to the receiving system. 1.6.4 .Describe any reporting capabilities of the system. 1.7 Solution must be capable of extracting additional data elements from the submitted document, such as student name, birthdate (even partial if only month/day are provided), graduation date, etc., if those data elements exist on the document. 1.8 Solution should allow for an unlimited number of licenses. 1.9 Describe the licensing model and how usage is billed. Are there any restrictions on volume of transcripts or documents processed?
