QUALITY ASSURANCE

99.5%: Handwritten, Cursive or Very Old Documents, Unclear Tiff images
99.995%: Typed or Printed Documents. Remote updating of web sites

SOFTWARE

We code from images using internally developed image entry software.
For PDF files our own developed process.
To create OCR output, we use only OCR Plus!

STEPS

The basics of the case need to be understood by the coders in order to identify and generate meaningful indices for the documents. Coders are trained as to case information as appropriate by the Project Manager; key information about the coding convention specifics of this case is also documented in the coding manual.

This information is also important in understanding the accessing needs of the end users. This will help in generating “Enhanced” or “Derived” Titles wherever “Natural Titles” are not present on the documents.

Defining the document boundaries or unitization is the most crucial activity in the entire coding process. This is done by Senior staff specialized for this activity and who are knowledgeable about the case contents.

This helps avoiding intermixing of coded information across documents. We then restrict coders to specific fields of the now “unitized” documents.
Different characteristics (fields) of the documents that are usually coded in our CodingPlus service :

1. Date
2. Document Type (Client selects types used in case from over 200 types)
3. Document Characteristics (e.g. Confidential, Redacted, handwriting, etc)
4. Subject / Title (or)
5. Subject / Title (enhanced)
6. Author/ Organization
7. Recipient/ Organization
8. CC/ Organization
9. And so on based on the individual case.

CODING PHILOSOPHY

Allow the coder to concentrate and get Quality Output.
Code one characteristic of a document at a time. This allows us to develop the expertise in locating the fields as well as improves quality and throughput.
Require complex fields like Doc Title to be coded by senior coders.
Two Data Entry Cycles followed by One QC cycle for 99.5% accuracy
For PDF files: Tiff-Clean, OCR, Char-by-Char check, Tables buildup, Image insertion for 100% accuracy.
Validation of date fields using normal algorithms for date validations.
Verification of doc type versus the approved list of document types for spelling mistakes and unwanted / undefined doc types.
Generating a master file of Author / Recipients / CCs on a continuous runtime basis and use the same at the final stage to verify the spelling as well as save runtime coding using drop down menus.
Checking continuity of Image and Bates Numbers.
Verifying Document Boundaries and Attachment Boundaries against the databases generated at the time of scanning so that integrity of the sequence is maintained and attachments are not intermixed.
Output is generated as specified by our Client

QC MEASURES

Validation of date fields using normal algorithms for date validations.
Verification of doc type versus the approved list of doctypes for spelling mistakes and unwanted / undefined doc types.
Generating a master file of Author / Recipients / CCs on a continuous runtime basis and use the same at the final stage to verify the spelling as well as save runtime coding using drop down menus.
Checking continuity of Image and Bates Numbers.
Verifying Document Boundaries and Attachment Boundaries against the databases generated at the time of scanning so that integrity of the sequence is maintained and attachments are not intermixed.

CODED DATA VALIDATION

Quality is an integral part of each and every function in our process execution.
Our Quality Assurance & Testing group is a central resource, which gets involved with the implementation team at the early stage of the process life cycle.

Date Validations:

The date is validated according to the format given by the PM with the following rules:

1. Month should be between 1 and 12
2. Day According to Month Name and Leap year (e.g. January has a maximum of 31 days or June maximum of 30 days.)
3. Slash Positions and Length: The slash positions of a date vary according to the format. For example it can be MM/DD/YYYY or YYYY/MM/DD or there may be no slash such as for date format YYYYMMDD. For European documents the format may be different.

Document Type:

Validations for document types are with respect to list provided by the client. All document types should be consistent to that list.

1. Spell Check
2. Case of the document type

Names:

1. Spell Check
2. Check if name gets repeated in same record but in different fields or in same field at a same record
3. Repeated Names

Title:

Spell check is done according to the MS Word dictionary.

1. Spell Check
2. Check case used for title

Characteristics:

Characteristics of a document type are also checked according to the list provided by the client and no characteristics are allowed to be keyed which is not present in the list.

1. Spell Check
2. Validate from given list

HIGHLIGHTS OF OUR CODING PRACTICE

Code directly from images.
Unique coding process to assure maximum accuracy of data.
Experts to handle logical unitization.
Coders have a minimum of Bachelors Degree with fluency in English.
Coding software can be easily customized to suit project specific needs.
Automatic data validation.
100% Quality Control measures.
Coded files are encrypted and zipped before uploading to client.

INFORMATION SECURITY –PERSONNEL RELATED

Access control systems in all premises include 24 hour uniformed security, proximity card readers & employee photo ID cards.
Access to the internet, email, printers and media drives is strictly controlled.
Zero paper policy provides tighter information security.
We conduct background checks for all new employees and have confidentiality agreements in place.
The work area for certain client teams are isolated to maintain security.
Client details and other confidential project information are not disclosed to employees or used internally.

INFORMATION SECURITY - TECHNOLOGY RELATED

Domain based network, access control dependent on functional needs
LAN & WAN completely protected through frequently updated Firewalls and Antivirus Software
Automated backup and recovery in dual media
Qualified technical network administrators conduct periodic assessments of technology resources to ensure environment integrity
Inventory control systems ensure receipt and processing of every received image, ensuring data protection and output qualit