Kira Systems: what if all the heavy lifting could be done by AI?

Today we asked Jorrit Berkhout, Account Director EMEA at Kira Systems, to tell us about Kira, and how it has revolutionized the legal profession. Kira is a machine-learning software that identifies, extracts, and analyzes content in contracts and documents with unparalleled accuracy and efficiency.

What Kira is? How does it work?

Kira is a machine learning software that identifies, analyses and extracts content in contracts and other documents.

Kira is a SaaS (Software as a Solution) platform that focusses on an easy-to-use experience. Lawyers have to simply upload documents that they want review with the help of Kira and will benefit from an automated process straight away.

Professionals in the legal space traditionally use Kira for due diligence activities in the M&A and Real Estate space, but we are observing an increased interest for firm wide adoption by using Kira in support of Finance, Compliance and Knowledge Management use cases.

From increasing the efficiency of the contract review process to the flexibility of building and integrating solutions together, Kira allows law firm customers to easily extract answers and insights from their contracts. Access to Kira comes with over 1.000 machine learning models included. These models help law firms to efficiently and accurately extract common clauses, provisions, and data points. Effectively providing customers to quickly get started on contract review projects and focus on the value-added activities for clients.

Also, each of Kira’s customers has the ability to easily create their own models to find specific clauses to meet the unique needs of each client.

Something that really stands out, are our customer facing teams who receive positive feedback from customers on a daily basis. Our technical (support) teams and practice consultants work hard to answer questions, meet every of our customers’ needs and to enable law firms to use Kira in the best possible way.

Why adopt a legal tech solution such as Kira in a law firm?

Professionals in the legal community can use Kira to gain time efficiencies of up to 80% in their contract review process, allowing lawyers to focus their time on more valuable work for clients, such as deeper due diligence and strategic legal consulting. This can prompt upsell opportunities to gain additional streams of revenue. Firms can also conduct a more comprehensive review of all their contracts to reduce errors by identifying and mitigating risks quickly, helping clients ensure compliance and avoid penalties.

In addition, these efficiency improvements can lead to a potential reduction in employee churn by improving the quality of lawyers’ professional lives. No one goes to law school dreaming about conducting due diligence.

How does Kira accelerate the lease summary process? And how accurate it is?

Kira uses supervised machine learning, meaning a lawyer feeds provisions from real contracts into a software designed to learn from those examples. Kira studies this data, learns what language is relevant, and builds provision models. The models are then tested against a set of annotated agreements that the software is unfamiliar with in order to determine its readiness. This highly accurate machine learning technology can identify and analyse virtually any provision in any contract.

The exact accuracy, as with any technology, depends on a combination of factors and one of them being the number of examples provided. One of the reasons that Kira is working so well for our law firm customer is that the software provides full transparency around the models such as precision and recall scores that help with managing what to expect from Kira in support of each project.

What our customers appreciate is that the end-result for their clients depends on Kira and the lawyers working well together, instead of either a fully automated- or manual process.

The increased productivity, as a result of the lawyer and software working together, helps our law firm customers by increasing their realization rates, gives them more opportunity to grow their revenue and preserve their existing clients.

What are the limits of Kira to date? In other words, which parts of a lawyer job becomes completely automatized and which are still not? What are the prospects for the near future?

Early perception of AI in the legal industry was that lawyers were going to be replaced by computers. But if you take the time to understand the tools’ nuances it becomes more and more apparent that AI enhances and lends valuable context to the work lawyers do rather than replacing it.

AI tools help legal teams reduce costs and increase productivity by automating certain components of repetitive tasks and processes long known to consume lawyers’ valuable time and cost the business money. In both transactional due diligence and contract life-cycle management (CLM), for example, AI tools enable lawyers to automatically identify and collect specific clauses among thousands, even millions, of documents within minutes. This means that firms and organizations that invest in AI tools like Kira will have a competitive advantage by letting their lawyers focus their subject-matter expertise on the most important issues.

More recently, beyond just automating contract processes, the next generation of contract analysis technology (including Kira’s Answers & Insights) is being used to extract and leverage “contract intelligence.” That refers to the critical business information embedded in contracts that has value to many business functions beyond legal, and for us, it represents the next frontier of AI-driven contract technology.

How does Kira mitigate the risk of errors and unexpected liabilities?

Kira’s standards require virtually every smart field to achieve a minimum of 90% recall, meaning the software will find 90% or more of the provision, clause, or data point you’re specifically looking for within your contracts or documents. Customers can also manage the accuracy and performance of custom fields they build within Kira, by adjusting precision, recall, and bias to optimise results. By training their own fields, customers also achieve a level of accuracy that might best reflect their own understanding of a contract provision, and they can adjust and manage the accuracy and performance of those fields as needed. 

Simply getting a more accurate picture of what is in your contracts is the best way to mitigate risks. Think of Kira as a tool that scales the lawyers’ ability to do what they are already doing. Instead of only reading a sample set of contracts in a database of contracts, the software can read them all. Instead of relying on the ability of the lawyer to manually read every example of a certain clause, he or she can extract every occurrence of that clause and review them in a structured report table.

Using machine learning mitigates some of the risks that humans bring to their work in large-scale reviews; stress, lack of time, lack of focus, lack of stamina. The human review is still necessary, but when lawyers use Kira, the software takes the load of the repetitive clause extraction off their plate, leaving more time for the human analysis that is the heart of the value they provide.

What kind of built-in provision models does Kira have? And considering it uses machine-learning, how fast, adaptable and reliable is the “Kira Quick Study”?

Kira contains more than 1.000 out-of-the-box provision models and this number grows consistently as we keep building new ones.

These so called ‘smart fields’ cover a variety of different domains, including M&A due diligence, Real Estate, Lease Accounting, Credit Agreements, Data Privacy and many more.

To the extent that our customers are looking for provisions, clauses and data points unique to their organizations or in different languages, they can create their own smart fields using the exact same machine learning technology behind our built-in smart fields with Quick Study. Kira Quick Study allows subject matter experts to capture information in contract according to their own expertise and to create their own fields that efficiently find specific clauses, provisions, data points and help answer their most pressing questions.

Kira can read and identify any Latin-based languages, so teams are empowered to meet the unique needs of clients or businesses with no developer skills needed.

The smart fields are also built with our powerful supervised machine learning technology, aimed to deliver results with the highest accuracy standards not previously possible with traditional rules-based systems. With a pure machine learning approach, the idea is that we teach the models once and our users can leverage those models over and over again. Kira calculates accuracy statistics for every smart field that users train. Our customers can manage and adjust the precision, recall, and bias of custom smart fields to optimise results and meet specific needs. Quick Study fields can be refined at any point by training with additional documents and document types to further improve the accuracy and expand the relevance of those fields.

How much does it take for a non-Kira user to get completely fluent with the platform? Do you think that lawyers’ skills are today in line with such innovative tools like Kira?

New users to Kira can get started with their projects immediately, thanks to our focus on ease of use and the intuitive user interface. Our customers – majority of them lawyers -have reported that:

  • Kira has the most user-friendly AI platform they’ve seen;
  • Kira scored highest on the ease of use metrics of all the products they compared;
  • anyone can use it with a minimal amount of training.

Included within every Kira subscription license are services that we provide around adoption, daily support and training. Customers can ask for our assistance whenever they want, and/or via a structured approach, in order to make use of Kira in the most successful way.

We also realise that every law firm works different and the software has proven so flexibility that every (unique) customer will recognise what happens on the screen in support of their client activities.

Does Kira use external servers or out-sourcing server management? If so, what are the safety protocols it takes into account? Are any data saved and if so, what are their uses besides training the machine-learning algorithms?

Kira uses Amazon Web Services (AWS) to provide our systems cloud infrastructure. We use industry best practices to secure our environments and each Kira customer cloud instance is contained within a distinct Virtual Private Cloud (VPC), ensuring that no resources used with that instance are shared with another. 

Kira Standard Clouds and Private Clouds have the same architecture, but differ in the number of customers instantiated in each; Standard Clouds are multi-tenant, with resources shared between those tenants, and Private Clouds are single-tenant, with a particular customer (or a limited group of customers) being the only tenant(s). These two Kira Cloud instance types do not differ technically. In multi-tenant clouds, a separate and distinct database is provisioned for each customer.

We don’t save any data uploaded by customers for our own purposes at all – the only use made of customer-uploaded data is for customers to train their own machine-learning models and for our primary use-case.

About Jorrit Berkhout

Jorrit is responsible for Kira’s client relationships in continental Europe and for expanding Kira’s presence around the world. In his role for Kira, Jorrit helps law firms to determine what activities and workflows can be automated via the use of software. He helps law firms to stay competitive via innovation of client services and improving their strategic approach. He focuses on being pragmatic and is known for challenging the status quo in order to push people to (re-)think what they are doing on a regular basis and how to improve this by implementing change. Jorrit worked on the digital transformation of clients via various commercial roles within the computer software industry for over 10 years. Clients that Jorrit impacted in positive ways via process improvement are active in a wide array of sectors, such as Financial (Banking)-, Legal- and Public Services, Oil & Gas, Automotive, Aviation and Technology.

For further information see:

Se vuoi dirci la tua e rimanere aggiornato sul mondo legaltech unisciti alla community Telegram! Clicca qui per partecipare!

Lascia un commento

Il tuo indirizzo email non sarà pubblicato. I campi obbligatori sono contrassegnati *