OpenSource Sanitize Can Speed Up KM

Just like many other knowledge organizations, we at Pillsbury Winthrop LLP have invested thousands of dollars in processes, training and technology to capture and leverage our partners' knowledge. Early on we realized that the ability to share personal knowledge and prior work would determine our place in a very competitive market. Knowledge Management had the potential to flush out communication and re-use issues, create a cohesive firm identity and ensure consistency in information quality.

With this in mind, 2 years ago, we set out on a comprehensive KM strategy that would ultimately improve how we manage firm information. Identifying and implementing the tools and processes needed to best manage electronic client, matter and expertise files took some time. Some of the tools were readily available, but others, including Sanitize by OpenSource, took significant time to either find or develop.

Once we began using our knowledge files to manage expertise, we found that most, if not all, documents created as part of a transaction include sensitive information pertinent to that transaction. This data can include deal details, such as price; company identifiers, such as name and address, and personal information, such as individuals' names, titles, and phone numbers. Similar to the metadata of a shared file, this data can give away potentially sensitive information. Removing sensitive details manually is very time-consuming, and can lead to errors and inadvertent information exposure.

A solution to this issue was simply impossible to find. We turned to the OpenSource folks, who specialize in automatic understanding of legal text, and whose products we use in other areas. They listened to our needs and developed Sanitize. Although there is no system that can provide a 100% accurate review of all sensitive information in a document, Sanitize gives our lawyers a powerful yet easy to use tool for reviewing prior work and removing potentially compromising information that might otherwise go undetected.

How It Works

The OpenSource Sanitize module is fully integrated with MS-Word, the main drafting tool that our attorneys use. Just like our metadata cleaning tool, a VBA agent (Word macro) automatically invokes Sanitize, whenever a document is filed into our knowledge folders. Sanitize can also be launched via a toolbar button.

OpenSource Sanitize analyzes the text of the document and removes pre-defined data. In our case, for transaction agreements this includes prices, names of companies and people, addresses and telephone numbers. This information is substituted with dummy data of our choice. The whole process is done in the background, without the intervention of the attorney.

Before finalizing the document, Sanitize requires feedback from the user, in our case an attorney. Currently, the system is setup so that it may capture information that is not necessarily confidential (false positives) but would not omit information that may be sensitive (no false negatives).

The user can save his/her changes and return to the feedback screen later, or “finalize” the document. Finalizing the document applies all the required substitution changes and invokes the VBA agent, which then returns the document to the Pillsbury Winthrop network in Word format. The document is now safe to post in our Knowledge folders, and otherwise share across and outside the firm.

Conclusion

OpenSource Sanitize is a small but crucial component of our Knowledge Management system. It enables our attorneys to trust the system, knowing that it operates within the confidentiality standards we are committed to.

The product is young ' it was created based on our request just a few months back ' and still needs to be expanded to support additional document types. Notwithstanding, I would recommend it to anyone who is trying to implement a Knowledge Management effort. As far as I know, this is the only working solution available, and it gets better and better with every release.

Just like many other knowledge organizations, we at Pillsbury Winthrop LLP have invested thousands of dollars in processes, training and technology to capture and leverage our partners' knowledge. Early on we realized that the ability to share personal knowledge and prior work would determine our place in a very competitive market. Knowledge Management had the potential to flush out communication and re-use issues, create a cohesive firm identity and ensure consistency in information quality.

With this in mind, 2 years ago, we set out on a comprehensive KM strategy that would ultimately improve how we manage firm information. Identifying and implementing the tools and processes needed to best manage electronic client, matter and expertise files took some time. Some of the tools were readily available, but others, including Sanitize by OpenSource, took significant time to either find or develop.

Once we began using our knowledge files to manage expertise, we found that most, if not all, documents created as part of a transaction include sensitive information pertinent to that transaction. This data can include deal details, such as price; company identifiers, such as name and address, and personal information, such as individuals' names, titles, and phone numbers. Similar to the metadata of a shared file, this data can give away potentially sensitive information. Removing sensitive details manually is very time-consuming, and can lead to errors and inadvertent information exposure.

A solution to this issue was simply impossible to find. We turned to the OpenSource folks, who specialize in automatic understanding of legal text, and whose products we use in other areas. They listened to our needs and developed Sanitize. Although there is no system that can provide a 100% accurate review of all sensitive information in a document, Sanitize gives our lawyers a powerful yet easy to use tool for reviewing prior work and removing potentially compromising information that might otherwise go undetected.

How It Works

The OpenSource Sanitize module is fully integrated with MS-Word, the main drafting tool that our attorneys use. Just like our metadata cleaning tool, a VBA agent (Word macro) automatically invokes Sanitize, whenever a document is filed into our knowledge folders. Sanitize can also be launched via a toolbar button.

OpenSource Sanitize analyzes the text of the document and removes pre-defined data. In our case, for transaction agreements this includes prices, names of companies and people, addresses and telephone numbers. This information is substituted with dummy data of our choice. The whole process is done in the background, without the intervention of the attorney.

Before finalizing the document, Sanitize requires feedback from the user, in our case an attorney. Currently, the system is setup so that it may capture information that is not necessarily confidential (false positives) but would not omit information that may be sensitive (no false negatives).

The user can save his/her changes and return to the feedback screen later, or “finalize” the document. Finalizing the document applies all the required substitution changes and invokes the VBA agent, which then returns the document to the Pillsbury Winthrop network in Word format. The document is now safe to post in our Knowledge folders, and otherwise share across and outside the firm.

Conclusion

OpenSource Sanitize is a small but crucial component of our Knowledge Management system. It enables our attorneys to trust the system, knowing that it operates within the confidentiality standards we are committed to.

The product is young ' it was created based on our request just a few months back ' and still needs to be expanded to support additional document types. Notwithstanding, I would recommend it to anyone who is trying to implement a Knowledge Management effort. As far as I know, this is the only working solution available, and it gets better and better with every release.