The usually calm world of legal and compliance suddenly finds itself at the forefront of change - and it’s not an altogether comfortable experience.
“Our legal department was slow to adapt when the company completely shifted its growth focus”, admits the General Counsel of a telecommunications company in a recent Gartner note.
“It hurt legal’s credibility, and mine personally.”
Considering, as Gartner goes on to estimate, that “…the average legal department lawyer spends 25% – 40% of their time on work that doesn’t even need to be done by a lawyer — at an annual cost of $2.7 million”, then it’s unsurprising that 80% of legal and compliance departments consider digital transformation as a must-do.
Data analytics and Artificial Intelligence (AI) look like the most likely candidate technologies to achieve this, with 16.7% and 53.3% respectively of law firms already testing them.
The sudden flip in expectations is mostly down to the acknowledgment that data analytics can now extract meaning, rather than just data, from documents, and that AI is capable of initiating action based on that meaning.
The need for decentralized AI
There are obstacles on the path, of course.
One is the resistance to sharing confidential data with other firms, something that is necessary in order to have bigger dataset to better train the needed algorithms.
Another is the lack of transparency of some common algorithms used in the legal world (i.e. deep neural networks). It is fear of this kind that lies behind the recent decision of the French Government to ban data analytics related to legal rulings.
Fortunately, a potential remedy is already at hand, in the form of ‘decentralized AI’, where processing and analysis of data takes place at the edge, under the control and visibility of users who understand the necessary regulations, professional codes and ethical concerns.
Fujitsu, for example, has developed Sholark, an advanced analytics and AI framework which can be a game-changer by combining its capacity to industrialize AI solutions together with decentralized AI technologies.
It has two variants targeted at specific users - Sholark Legal & Compliance, and Sholark Health - each adapted to the complexities of their respective sectors and guided by the knowledge and experience of expert users.
Reaching end users
To be fully practical, an AI system also has to be accessible outside the IT department. It must be usable by the sector specialists – lawyers and paralegals in this case – who fully understand the constraints of their profession.
Sholark is designed to be easy to use for non-technical end-users and Fujitsu has put aside any ‘black box’ paradigm and given non-technical users the ability to maintain control in data analytics processes.
The Sholark Legal & Compliance solution allows non-technical users to address four of the key business challenges facing departments across all industry sectors.
- Data visibility. Availability and visibility of information - or the lack of it - is one of the biggest causes of inefficiency in legal and compliance. Using Natural Language Processing (NLP) a ‘knowledge graph’ can be created about the information in legal documents, finding all the links between the items of information, and deducing non-obvious, hidden data links, as when legislation has resulted in positive outcomes in two apparently unrelated cases.
- Data analysis. Advanced analytics also make it possible to compare different cases and to make predictions. In litigation, more accurate predictions of the likelihood of winning a case could help companies decide whether and how to settle. This sort of risk management often goes by the wayside, even in high-value work, with only 15% of legal departments integrating legal risk requirements into business workflows, and only 21% tracking the efficacy and impact of risk mitigation plans.
- Automation. Lawyers’ time is expensive. Any mismatch between the complexity of a task and the seniority of the person conducting it is a recipe for inefficiency. Automatic indexation of legal documents as they are received via email allows for a high degree of automation. The AI reads and understands documents, making automatic summaries so users can quickly decide whether it is important or not. It can also automatically classify documents, extracting and categorizing judicial court sentences, for example, or identifying important clauses within a real estate contract.
- Knowledge sharing. Knowledge sharing is often weak or wholly absent, as legal experts’ knowledge is usually contained within their own heads or PCs. This tends to lead to decision-making based on the experience and knowledge of individual users, who might only be aware of a limited set of documents. Knowledge graphs digitalize the experience of different users, opening up access to knowledge and jurisprudence contained within the wider group.
Out in the real world
If this sounds like some unattainable nirvana, Sholark has already solved various legal and compliance challenges that often create day-to-day bottlenecks for lawyers, such as automatic classification, search and summarization of judicial sentences.
For a Spanish Government ministry, Sholark has reduced the time dedicated to this type of task by 60% and increased the efficiency of its cataloging process. Its cognitive computing techniques are generating automatic document summaries to simplify the review process and tracking the performance of relevant KPIs, such as the number of sentences per organization or per category, the duration of the legal process, most common searches, etc.
With automation in legal & compliance now a viable option, the department pain described by Gartner is now optional.