Artificial Intelligence

The term AI is used to refer to a number of different technologies, which exhibit functionality that we associate with human intelligence.

AI has evolved to use significant computer processing power to combine and analyse vast quantities of data and documents, together with an ability to understand natural language.

Legal AI systems are machine learning technologies that apply pattern recognition in training data (often human-analysed legal documents) across much larger data sets. These technologies can be “taught” new patterns on a continuing basis, through significant data input and the application of legal rules and logic, transforming our ability to search, process and analyse large volumes of documentation. As use and legal data multiplies, AI-based tools enable more sophisticated data analytics that have the potential to complement and inform the advice that we provide to clients.

One example of how we can use these tools to support our clients is the ability to automate repetitive, and often arduous, tasks such as the document review component of a due diligence exercise. In addition, by combining our legal excellence and cutting edge technology we can provide a distinct and compelling competitive advantage to our clients. We are continually developing innovative technology solutions in order to meet our client’s individual evolving requirements. Through integrating this advanced legal technology into our existing capabilities we aim to provide a leading service for our clients.

What does it offer our clients?

  • Time and cost savings – through utilising these technologies lawyer time can be reduced on large reviews which can be dedicated to more strategic or high-value areas.
  • Increases consistency and accuracy – loss of “element of human error” can improve the quality of service delivery.
  • Quicker evaluation – Legal AI can perform analysis on extracted data and answer specific questions, reducing the overall time it takes to evaluate certain information. It can provide “yes” or “no” answers and is able to query legal provisions on transferability, assignment and mortgages, among others.
  • Aiding decision making in “real time” – on more routine matters eg when negotiating terms “in the field” Legal AI can identify non-compliance by percentage-ranges with standard terms for documents.
  • Bespoke solution – adapting technologies to provide unique packages specifically for our clients’ individual requirements.
  • Better workflow management – meaning more predictable action times and therefore more time and cost control.
  • Agile technologies – products which can be used on tablets or other forms of computing device meaning more versatile and immediate deployment.
Where can Legal AI be applied?

Legal AI can be used to improve efficiency on a broad range of tasks and on a variety of document types. These tasks might include (and are not limited to):

  • Data Room Management: Legal AI technology can make the collaboration of a data room quicker and more cost-effective as a result of reduced lawyer and paralegal time. Legal AI technologies can assist in cleaning up data rooms by flagging poor quality scans, duplicate documents, unsigned agreements and missing pages. The systems are also able to cluster documents into groups (for example by similar document type or size), which supports the identification and prioritisation of information.
  • Due Diligence: Legal AI systems can enable a reviewer to efficiently analyse significant numbers of documents. They automatically extract or answer questions relating to key data points, reducing the time taken to identify important information at the initial stages of review. For example, Legal AI can be used to review a data room for issues relating to change of control, conformity with internal standard forms or consent requirements. It is also possible to conduct advanced searches across selected documents and, when using the external tools, to review in the interface itself and assign tasks to individuals on the matter. The larger the document set, the greater the efficiencies to be gained.
  • Workflow functionality: Legal AI systems have the potential to help workflow management on matters. The systems allow project owners to set permissions and allocate tasks within the tool, supporting a more efficient, streamlined process.
  • Disclosure: AI functionality embedded within ECA tools can be utilised during the initial stages of a disclosure exercise to uncover connections between documents and extract and relate concepts, as well as de-duplicate and cluster at a rate of nearly 2 million documents per hour. Technology assisted review can then be leveraged to complete the review of these documents in the most efficient manner, saving both time and costs. The system “learns” from the review calls made by a subject matter expert as to whether documents are responsive or non-responsive and this learning is then applied to the wider document population in order to quickly identify highly responsive documents. The use of machine learning has been advantageous in terms of both accuracy and speed when conducting reviews for the purposes of disclosure.
  • Document Comparisons: the comparison functionality of the Legal AI technologies potentially enables a more efficient evaluation of mark-ups of simple documents (eg NDAs) by automatically undertaking document comparison and highlighting key areas of difference for review by lawyers.
  • Delivering complex legal expertise: Legal AI tools can also be used to significantly simplify access to complex legal analysis, such as in the regulatory field. This works by combining our in depth and often multi-jurisdictional legal expertise with underlying algorithms and decision trees. Bespoke tools for particular areas of practice can then be created and made available both to our clients and our internal teams to drive both quality of analysis and efficiency of delivery.

We are at the forefront of Legal AI programmes

We use a dual approach when applying Legal AI technologies

Each tool has different strengths and requirements and will require time and financial investment to train them to achieve the best results. They can be used for a variety of functions, including data clustering, data extraction and data analytics but there is no single solution to all of our client needs.

External solutions

These systems automatically read, categorise and extract information, which is then condensed into a specific template for our team to analyse. Our lawyers then “teach” the system where provisions are identified incorrectly. Products we have access to globally include RAVN, Kira, Leverton and Neota among others. Linklaters was the first magic circle law firm to sign with RAVN and has developed a deep understanding of the tool’s capability through several rounds of testing.