When we think about artificial intelligence, what comes to mind might be HAL 9000 from 2001: A Space Odyssey, Skynet from Terminator, and even Ultron from Marvel (a.k.a. Tony Stark’s mad AI experiment). You know, robots are going to take over the world. Humanity is doomed. Computers for life. The matrix is real…
Thankfully, not all AIs are so destructive and dangerous. For every Terminator, there’s also R2-D2, WALL-E, and Data from Star Trek. Suffice to say, Earth is safe for now, and chances are low that we’ll find ourselves in a world where machines take over and force humans to fight as an underground resistance.
Most of the AI we see in films and series are still works of fiction. Whether we get to that level, only time can tell. But be that as it may, there are many advanced applications of AI in everyday life, including AI in contract management.
Before we dig deeper into how AI plays a role in CLM, here are a couple of key definitions.
What is Artificial Intelligence?
AI attempts to replicate humans’ intelligence or behavioral patterns. Theoretically, an AI can be created based on some other living entity, so long as its intelligence or behavior can be mapped, but for simplicity, let’s assume all AI in legal technology tries to replicate human behavior.
Creating an AI requires the application of techniques that enable computers to solve problems as if it were a human. This usually takes the form of a set of pre-defined rules.
What is Machine Learning?
Within the larger sphere of AI, ML is a field comprised of four subcategories: supervised learning, unsupervised learning, reinforced learning, and deep learning. It’s a technique that allows computers to “learn” from data. In other words, it’s a type of AI that learns by itself.
Often, ML is conducted by “training” a computational model (also known as an algorithm) using a dataset. The more time and data provided, the better the computer’s performance.
A subset of machine learning is deep learning. This attempts to use artificial neural networks, which are computer systems that imitate the human brain’s neural network. These ANNs are designed so that an increasing number of characteristics are extracted from the input data.
In simpler terms, deep learning is simply machine learning, with the key difference being that the model is based on human neural networks.
Machine Learning vs Artificial Intelligence
Some people use AI and ML interchangeably, as if they were perfect synonyms. Others might refer to them as completely separate, parallel examples of advanced technology. Sometimes this is because both words are overhyped, and sometimes it’s because people don’t quite get the difference.
The short answer is that AI and ML are not the same.
As we mentioned earlier, machine learning is a field of AI. They’re closely related, yet still different. All machine learning counts as artificial intelligence, but not all AI counts as machine learning. AI, in fact, can be divided into four primary parts: Reasoning, Natural Language Processing (NLP), Planning, and Machine Learning.
AI can be divided into four primary parts: Reasoning, Natural Language Processing (NLP), Planning, and Machine Learning.
An example of an AI that doesn’t use ML is a chatbot. Certainly, there are chatbots that use ML, but a basic chatbot doesn’t require it. That’s because it’s a rule-based system whose “rules” (the questions it asks and answers it receives) were defined by people. These sorts of “expert systems” are essentially a sequence of “if this, then that” statements, otherwise known as a decision tree. Although no learning occurs, such rule-based chatbots can still prove highly useful.
But let’s imagine we decide to create an ML-based chatbot. Whereas all questions and answers were previously scripted for a rule-based bot, an ML-based bot is given a massive corpus containing hundreds of thousands of conversations to clean up and analyze. Then based on the information provided, the ML-based bot is trained to handle various situations.
AI and ML in everyday life
Most likely, you interact with some tool or program that uses AI or ML, such as any Google product. Major services like Gmail, Google Search, and Google Maps all have ML. For example, how emails are filtered into the primary, social, and promotional tabs, or how Search and Maps anticipate what you’re looking for based on past searches, trends, and even your location.
Then there are assistants like Siri and Alexa, social media newsfeeds, Netflix, Amazon, and even smart home devices. As the world becomes increasingly digital and smart functionalities become more ubiquitous, it’s almost certain that you’ll be interacting with ML-based products and services on an hourly basis.
AI and ML techniques in contract management solutions
To quickly recap AI and ML, think of artificial intelligence as an attempt to give machines the intelligence of people. In contract management, this would enable an AI to read and interpret agreements while extracting key information.
Machine Learning is the process of teaching or training the AI by exposing it to massive quantities of documents. The larger the dataset, the better the AI’s accuracy and the more stable its performance.
But one critical technique that hasn’t been highlighted is NLP, which is the ability for software to recognize and mimic a person’s speech. This approach helps identify and resolve potentially trick spots in the language. It’s also used when contracts are being drafted, when spotting missing contract aspects, and categorizing contract attributes with regards to the context.
Role of AI in CLM
Currently, there are three core roles that artificial intelligence plays in contract lifecycle management:
Data extraction from active contracts
With the help of ML and Natural Language Processing, legacy contracts and other documents you had before implementing a CLM software system can be quickly imported to the new CLM platform. This is of critical importance if any of the contracts and agreements are still active. To avoid complacency and missed deadlines, it’s a good idea to upload legacy contracts so that the new CLM software can do its job of managing obligations and generating notifications.
Since this can be done using ML and NLP, the importing process takes considerably less time. Teams no longer need to open PDFs, read terms, save them in Excel, and manually track them. That’s because both ML and NLP help identify a contract’s metadata, such as the due dates, deliverables, and signing parties. Better yet, technology is good enough that even if a contract is a scanned image (this might be the case if your signed contract is only in paper form), information can still be extracted.
This kills two birds with one stone because at the same time that you’re importing legacy contracts to the new CLM platform, you’re training the ML algorithm on your contracts. That means the ML will be more capable of handling your future contracts.
Assistance during contract authoring
During the stone age of CLM software, processes were simply made available in a digital format. For example, the system would simply provide fields that could be filled in with data.
CLM software has come a long way since that, in part due to significant advancements in AI. Trained on thousands of contracts from several years, a quality AI-based CLM software can act like a senior advisor by reviewing contracts and suggesting opportunities for negotiation and risk management. Assuming your organization has a clause library, it can even provide recommendations on which clauses to use according to geography, vendor type, and contract value.
Some CLM platforms even offer a chatbot for the contract authoring process. Not all are powered by ML, but they add a great deal of value for users by helping fill out contracts with the answers to select questions.
Management of contract obligations
Arguably AI’s biggest role in CLM, an AI can extract obligations automatically from contracts (including uploaded legacy documents) to manage them. At present, AI still has some problems when extracting non-quantifiable obligations like IP protection and employee welfare. But when the language features quantifiable terms like discounts and due dates, AI works like a charm.
Major applications of AI in contract management
As mentioned previously, AI has come a long way since its inception. The meaning and techniques used have consistently evolved over the years until we’ve become dependent on AI to simplify common problems. Google and Netflix are major examples of the strides that AI has made in solving peoples’ problems, but other industries like medicine and finance have also been affected. Legal is no different.
How an AI-based CLM can solve contract management issues
Here are just a few of the ways that AI can address common problems of contract management:
Batch review
Instead of checking one document at a time like a person would, an AI can review multiple agreements, update terms, and import legacy contracts in batches. For example, if your government regulator changes certain legal requirements, you can input the fix across multiple documents at the click of a button.
Risk
The more contracts your organization handles, the more likely there will be missed renewals, fee increases, and compliance problems. The failure to abide by contract terms is one of the leading causes of commercial legal disputes. An AI-based CLM is capable of generating the reminders you need and monitoring obligations so that your organization stays compliant.
Time
It’s been said before, and it’ll be said later, but the consequences of saving time are massive. Not only do you save money as a direct result, but your employees will have greater freedom to dedicate their efforts to more productive and professionally rewarding tasks.
AI can be used to help overcome challenges that frequently appear as contracts make their way through the workflow. But that’s not all. With the right approach, AI can unlock innovative, new possibilities.
For instance, GPT-3, which is a neural network model that uses deep learning to generate any type of text, opens up countless opportunities. Its ability to perform reading comprehension and writing at near-human levels stems in part from having consumed more text than any human can ever read in their lifetime.
Reasons to use AI in CLM
There are a lot of excellent reasons to use artificial intelligence in contract lifecycle management. But first and foremost, the top three reasons are that it saves a lot of time, a lot of money, and a lot of effort. Processes can be completed 80% quicker (which in turn saves money), and a lot of repetitive tasks can be carried out automatically.
Those are just some “big picture” reasons though. Here are some specific benefits that AI can provide legal companies and legal teams:
Automated contract compliance
An AI CLM system can be configured to take into account regulatory and contractual compliance terms that are industry or jurisdiction specific. Also, in addition to setting notifications for specific team members regarding dates and obligations, an entire audit trail can be created.
Maximized employee expertise
AI is unable to operate with complete independence. There must always be somebody overseeing its performance in order to guide and fine-tune it, or to step in if there’s an anomaly or problem. As a result, companies can leverage and maximize team members’ expertise while freeing them from repetitive, less interesting tasks.
Standardized contractual processes
AI can ensure that there’s consistency across the entire CLM process by putting every contract through the same sequence of stages. Not only is this the most efficient approach to managing contract lifecycles, but it enhances visibility from start to finish while simplifying the identification of any contracts with issues to be resolved.
Improved connectivity
Robust, organized systems can unlock countless opportunities for companies thanks to the increase in efficiency. AI contract software can clean up and standardize data so that it can be easily integrated with other powerful tech. Otherwise, if data remains inconsistent, and if storage and retrieval systems are outdated, then teams will miss out on the advantages that automated CLM solutions can provide.
Efficient scaling
If an AI-based CLM solution has made data standardized, compliance automated, and visibility ensured, teams can scale and add layers of management without becoming overly complex and inefficient. That’s because an AI CLM software system doesn’t need an increase in personnel to keep a growing organization running smoothly.
How AXDRAFT uses AI
AXDRAFT attempts to leverage artificial intelligence and machine learning technology to create the best contract automation experience for lawyers and legal teams. The CLM’s AI is constantly learning, and it enables users to craft contracts in any language, in any complexity, and at any time in moments. The fact it can generate 100 pages within seconds allows legal teams to spend 80% less time on drafting and proofreading while freeing up 40% of the work week for more productive projects.
And thanks to GPT-3, Legalese can quickly become a relic of the past. AXDRAFT’s Simplify feature, implemented using GPT-3, simplifies legal text into language that colleagues and clients with no legal background can comprehend. Similar to deep learning techniques, this approach supports three main uses: semantic search, classification, and text generation. And aside from leveraging this powerful tool to translate legal text to simple English, AXDRAFT deploys it to extract and classify data, as well as fill out contract templates.
With AXDRAFT’s help, several companies were able to speed up their processes and save valuable funds. For instance, Glovo began processing documents 92% faster, which led to savings upwards of $600,000 annually.
Conclusion
One of the most important things to remember about AI in contract management is that it’s not about to steal anyone’s job. Like we mentioned before, AI is best deployed in a supporting role where it can provide critical assistance in speeding up contract processes. Although modern AI has made great strides, it can’t function 100% independently. But if deployed correctly, it can speed up the contract workflow by up to 80%.
If you’re interested in how AI can help shape and enhance your organization, book a free demo with us at AXDRAFT. We’ll take you through the ins and outs of our CLM platform while demonstrating how we incorporated AI. Get your free trial here.
FAQ
AI contract management is designed to streamline the entire contract lifecycle. Various tasks such as data insertion, data extraction, data security, risk identification, and risk assessment can all be performed automatically by an AI. It can also help manage massive amounts of data generated by large contract volumes, apply standard operating procedures to all contracts, and share the load of administrative and specialist work.
As a result, less time, money, and effort is spent or wasted. Also, there’s less risk, more visibility, and greater opportunity. Team members spend less time on the mundane and more on possibilities that can generate positive revenue.
As a result, less time, money, and effort is spent or wasted. Also, there’s less risk, more visibility, and greater opportunity. Team members spend less time on the mundane and more on possibilities that can generate positive revenue.
Contract review is a process that identifies provisions, clarifies facts, assesses feasibility, and forecasts risks. During a review, legal professionals carefully read and analyze all aspects of a contract: terms, conditions, risks, and relevant information.
AI contract review does the same thing. The only difference is that instead of a human, the contract is reviewed by an AI that was trained on a set of hundreds or thousands of real contracts. AI can identify important contract features such as the type (sales, lease agreement, etc.), parties, dates, governing law, and more.
The AI can also trigger a certain workflow. For example, if your company has Workflow A for employment contracts and Workflow B for sales contracts, then once the AI learns that it has a sales contract, it will route the contract to Workflow B.
At the moment, the short answer is “it depends”, and the answer relies on your definition of the word “write”.
Many legal documents are being drafted with advanced AI technology since they can be produced faster with more grammatical correctness. AI is perhaps best utilized when the same contract templates are used day in and day out, with the only differences being specific data related to the counterparty (name, date, email, etc.). In these situations, the AI’s logic can be constructed so that if the correct data is inputted into the template, it can produce an error-free contract. However, this contract was based on a template or templates originally written by a person.
So while AI is not quite capable of writing contracts autonomously and completely from scratch (though some believe AI is very close), it can write contracts using templates and additional information. If you think about it, this is exactly what most businesses do. They use the same standardized contract templates over and over, changing only very specific details related to the agreement.
In the end, can it write a contract from scratch? Not yet. But can it write a contract using a template? Yes.
Still, a best practice would be to have a human quickly review the important details during the approval stage before sending the contract for execution. But this should happen even if a human wrote the contract. As they say, two eyes are better than one.
Just like the answer to whether an AI can write a contract, whether an AI can read a contract is also “it depends”.
In 2018, it was revealed that a study that pit a contract AI against 20 human lawyers from the United States ended in a victory for the AI. On average, the AI was not just 10% more accurate, it was much faster. The AI completed the task in 26 seconds while human lawyers took 92 minutes on average.
There are, however, certain aspects of language that can prove challenging for an AI to process. Although Natural Language Processing techniques are continuing to improve with each passing day, it’s still a good practice to have a human perform a double-check.
The important point to take away though is that AI can significantly expedite certain tasks. In an interview after the aforementioned study, having an AI perform the first review is like having a paralegal do the first review. This frees up valuable time for lawyers and paralegals that can be dedicated towards more productive, high-value work.