How we're using AI at Safelink

26 April 2024 Karl Anderson

When evaluating how new AI technology could be applied to Safelink’s products, it didn’t take long to determine that the best approach would be to enhance existing processes rather than replace them. Our products are mostly used by skilled professionals in legal and financial services, so respecting existing processes is essential.

We liken our approach to driver assistance in a car, rather than ones that completely take control. Even if automated driving were to work most of the time, a mistake could be devastating. Our goal is to improve the driving experience while allowing oversight of the processes, providing excellent controls and allowing the driver to reach their destination in the right way.

In this context, security is essential. Security is our foundation stone and our customers demand it. The risk of confidentiality loss, either through regular vulnerabilities, jurisdictional issues, accidental exposure to manual review, or incorporation of confidential information into AI training data all need to be addressed.

LLM technology in practice

We use Large Language Model (LLM) technology within the Safelink product suite in a few different ways. In some cases, we extract facts and metadata from unstructured information, to present those back to you in a structured way, such as in labels and metadata. In others, we're building full databases from documents, to give a simple solution to browsing facts, events, and parties.

AI opens the door to powerful filtering and search tools that aren’t usually available for unstructured data. There’s also the now-familiar use of concept indexing, which enables the use of chat and question-answering bots, fuzzy or concept search, and the ability to identify similar documents.

All of these features are designed to speed up existing tasks, particularly time-intensive ones, allowing you to focus on decision-making and strategic work.

How we use AI in our products

Chronologica turns documents into legal chronologies using AI

Here's how we've applied AI within our three products: Lexiti, Expero Virtual Data Rooms (VDRs) and Chronologica, to simplify operations:

Chronologica

The newest addition to the Safelink product suite is Chronologica, a seamless chronology-building software which automates the creation of case timelines. AI is used significantly throughout Chronologica with:

  • Automatic fact and event extraction from uploaded documents
  • Timeline generation, guided by the narrative you wish to cover
  • Suggestions of details and evidence for timeline events

Lexiti

Our popular eDiscovery tool helps to import, search, review and export documents. This streamlines narrative construction, so you can securely share data with your team. AI-assisted features within this product include:

  • Automatic document summarisation and naming of documents
  • Identification of key entities such as people, places and dates within evidence

Expero Virtual Data Rooms (VDRs)

Our Expero VDRs are used to share sensitive documents and have restrictive access controls to closely monitor all access. AI enhances our Expero VDRs with:

  • Concept search to identify documents containing similar concepts to your query
  • ‘Ask Alex,’ a question-answering bot, supported by document retrieval using the documents available within the data room, while also respecting the permissions applied to each document and folder

How can I enjoy AI benefits while ensuring data security?

Not everyone is going to want to use our AI features. We've addressed the inherent concerns of using AI, but the features are supplementary. Our products work well even with these AI features disabled; we’ll leave it to our you (or your IT team) to make the call. Expero and Lexiti customers can choose to opt in to AI features; for Chronologica, AI features are enabled by default. Don’t hesitate to contact our team if you’d like to discuss the use of AI in more detail.

Karl Anderson - Profile Picture

Karl Anderson

Founder at Safelink
Understanding the data security challenge When Safelink launched in 2010, our first customers were offshore wealth management firms....
Read More - How we built a system that houses some of the world's most sensitive data
On Monday, the US Federal Trade Commission (FTC) agreed a settlement of up to $700m for a class action brought against Equifax. This...
Read More - Learning points from the "entirely preventable" Equifax data breach
Thousands of British Airways, BBC and Boots employees have had their personal information stolen due to a weakness in file transfer...
Read More - How was British Airways payroll data stolen if it was "encrypted at rest"?