“Improving the DSAR headache” case study

Posted on 25/08/22

The Situation

A very large public sector customer of ours was, and remains, under pressure to maximise the amount of time they spend delivering front-line services, while working to very tight budgets and having to adhere to their environmental commitments.

 

The problem

In the first six months of the year the organisation received an increasing number of DSARs, many of which were of a complex nature. This change created several challenges, such as:

  • The data collection process, specifically where to store the data prior to review
  • The duplication of information, specifically how to efficiently find and discard it
  • Handwritten documents, specifically how to include them in the review process
  • Locating personally identifiable information, specifically for unknown and/or unexpected subjects
  • Reviewing the dataset, specifically when other (front-line) stakeholders need to be involved
  • Redaction, specifically how to speed up the redaction process across all files in the dataset
  • Defensibility, specifically how to ensure any redactions can’t be reversed in the disclosure
  • The final disclosure, specifically avoiding the need to print it off
  • Long term storage, specifically in the instance a future challenge is made

Overcoming the problem

Smartbox.ai was chosen as the technology solution of choice to help overcome the above problems, as it specifically enabled the organisation to:

  • Over collect and upload into the Smartbox.ai environment as many files and data repositories deemed necessary for each DSAR, regardless of their size or format. These consisted of exports of email accounts, enterprise social networks, and a range of documents – many of which were handwritten. Each time data was uploaded, Smartbox.ai began extracting and cataloguing the files, validating whether it considered the data to be sensitive, and producing high fidelity renditions of all information of interest.
  • Reduce the dataset by an average of 63% just be using Smartbox.ai to automatically identify, validate, and remove all duplicated information. In most cases the duplicated information was generated from email chains where information had been shared, forwarded, and replied to, creating multiple copies of the same information.
  • Include, treat, and rely on Smartbox.ai to automatically review handwritten and/or multilingual documents in the same way as digital documents. This enabled the amount of human effort needed to look through and redact handwritten documents to decrease by circa 90%.
  • Use the AI within Smartbox.ai to automatically find and highlight every reference to any information that may identify an individual, such as names, date of births, addresses, national insurance numbers etc. In addition, as the organisation had a unique format for identifying employees, Smartbox.ai’s “dictionary” feature allowed for the bespoke format, together with other relevant taxonomy’s to be entered into the system and used as part of the review and categorisation process. This gave the organisation the confidence that virtually every reference to a third-party individual would be highlighted without any human effort, thereby reducing the chances of sensitive data being disclosed accidentally.
  • Improve the review experience and in the process free up hours of time for front-line workers. This was achieved in a few ways. Firstly, by the 63% “deduplication” reduction in the dataset mentioned above which directly impacted the overall effort needed to review the remaining data. Secondly, by leveraging Smartbox.ai’s collaboration capability which allows multiple stakeholders to participate in the review process without having to take the data out of the Smartbox.ai environment; and thirdly, by Smartbox.ai’s advanced and ultra-fast ability to automatically identify all references to personally identifiable information, which directly reduces the effort needed to look through the dataset.
  • Redact information within the entire dataset with a click of a button. Much like other redaction solutions, Smartbox.ai allows documents to be reviewed and redacted one at a time, however, unlike other redaction solutions, it also allows every document in a complete dataset to be redacted with a single click of a button. This had a material time-saving impact for the organisation and improved the overall quality of the disclosures as there is now less chance of third party (sensitive) data remaining u-nredacted.
  • Have confidence in the defensibility of the disclosures. One of the concerns the organisation had was to do with the likelihood of redactions being reversed. With Smartbox.ai this is impossible as when approved for publication the redacted information is saved as a flat “rasterized” image, meaning it cannot be reversed. In addition to the rasterized version being produced, for audit purposes, an un-rasterized version is also produced, giving the organisation the opportunity to revisit cases should the need arise.
  • Reduce the amount they are printing. In general, there is virtually no reason why a DSAR disclosure needs to be printed off, and this is now the view of the organisation. Today, the organisation uses Smartbox.ai to securely share all prepared disclosures. When the disclosure is ready to be shared, a secure HTTPS link is generated which is sent to the subject. The subject then clicks the link and views the information, triggering an auditable read-receipt notification.
  • Store data securely and cost-effectively in case of a future challenge. Smartbox.ai’s long-term storage capability allows the organisation to save the redacted (and un0redacted) disclosures within the system. Should they be needed the disclosures are available at a touch of a button, giving the organisation peace of mind that should they need access to the information, it can retrieved within seconds.

To learn how Smartbox.ai may be able to help you, book a demo today.

< Back to News & Insights