In the remote work environment of COVID-19 we are likely to see increased automation of legal activities and processes.  In this post, we look at some of the hurdles to document automation including the process, knowledge engineering and technology, and how you can overcome these challenges to leverage opportunities or directives to further automate and streamline your legal work.  

Why Document Automation and Why Now?

While document automation projects can range from the very simple to the very complex, the basic premise is a questionnaire that is used to determine which fields to populate and the language to include.  At Stikeman Elliott, we use it in a range of situations such as repeat transactional work with specific clients, technical legal documentation where risk management is important and suites of ancillary documents for various matter types.

Document automation tools are not new but uptake in legal has been slow.  The tools have become more popular as legal departments and law firms look for efficiencies and adoption may be pushed further in the environment of remote work and COVID-19.  These projects are within your grasp whether you have document automation software or not, so we wanted to share some insights around the related process, knowledge engineering and technology.

Best Practices for the Document Automation Process

We were early adopters of Business Integrity's DealBuilder (now Thomson Reuters' Contract Express).  As a result, we have had time to hone our process and develop some best practices.  The process is documented in our Innovation Project Workflow which will work for law firms and law departments alike and is technology agnostic.  Each stage of the process can be its own hurdle needing tips and tricks to push through these potential roadblocks.

Initiation and project intake

"I have an idea."  That is how most automation projects start and while your lawyers and users may think they know what they need, it is important to listen to their use case and confirm that document automation is the right approach.  Sometimes it is not.

Precedent selection

It sounds obvious but it is a best practice to settle on the document and its variations before automation.  Based on our experience, if you skip this step you are essentially creating the precedent and doing the automation at the same time which means the project can take longer.

Knowledge engineering

This is the process of taking a document and deconstructing it to identify the variables and rules that will be needed to create the various outcomes.   See details on this key step below.

Coding

The easiest part of the document automation process! However, it does not mean that we have our mainstream lawyers coding templates.  We have found that it requires someone with knowledge of the underlying technology who uses it on a regular basis.  It is beneficial if they have legal knowledge as well and we have a mix of KM lawyers and KM specialists who support the coding required for our automation projects.

Test, review and revise

You need to develop some strategies for working with your lawyers or subject matter experts to get their input along the way.  Not all of our lawyers will test the automation on their own.  When that happens, the most effective way to move projects forward is to meet with them, show them the automation, gather feedback and revise.

Pilot and finalize

The key to finalizing an automated document is to use it with real examples.  As a result, we always include a "pilot" stage in our projects.

Communication, roll-out and training

Your approach will vary depending on whether you have a large or small number of people generating the documents in question.  We have found that it is important to train users on both the automation technology as well as the underlying document.

Maintenance

Finally, precedents and templates do not stand still.  Nor do rules and variables used in different circumstances.  Establish a process to gather feedback from users and make updates to the document including changes to the precedent and changes to the automation.

A Deeper Dive into Knowledge Engineering

This is probably the most intimidating part about document automation and is one of the big hurdles to successful projects.  You need a subject matter expert who knows the document intimately.  You also need to create very detailed instructions on how to code the document.  Bridging that gap can be difficult and is often where the automation process can slow down or stall.

Through trial and error, we have broken the process down into three steps and have developed some simple tools for our lawyers to use when they are assisting us with knowledge engineering.

Step 1: Identify Your Variables

A variable is a piece of text or information that will be entered into the document, usually in place of a bullet.  The first step is to go through the document and identify all of the variables and put them in [bold square brackets] and highlight them with a distinctive colour. E.g. [Corporation Name]

Step 2:  Identify your Rules

A rule dictates when certain text in the document will appear.  The second step in the knowledge engineering process is to go back through the document and identify the rules for the optional text in the document.  An easy way to do this is by using the comment function in Word.  For each rule, insert a comment that spans the text that is subject to the rule.  The comment should then be used to describe the rule with enough detail to provide instructions to the person who will be coding the document. E.g. "include this text if this is a two-way confidentiality agreement."

967654a.jpg

Step 3: Create Your Questions and Calculations

The third step is to create the questions and calculations that are needed for each variable and rule that you identified.  Questions are used if the information does not already exist; calculations are used based on answers to previous questions.  For example, if you have variables in a subscription agreement for (i) the number of shares being subscribed for, (ii) the price per share and (iii) the total consideration being paid, you would need questions for variables (i) and (ii) but you could use a calculation to determine variable (iii) based on the answers to (i) and (ii). 

We use an excel spreadsheet for this purpose that has columns for the information the coder will need.  This step is more likely to be technology specific, so keeping that in mind, the information we require is:

  • The name of the variable or rule;
  • Whether it is a question or calculation;
  • The type of answer - text, date, number, single select, multiple select, etc.;
  • The question to be asked and any options if a selection is to be made, or the calculation to be performed; and
  • Guidance, if any, that should be provided to the person answering the question.

Lawyers who work with us on these projects are provided with a step-by-step guide, a sample mark-up of a document and a corresponding sample excel spreadsheet. 

967654b.jpg

Your Document Automation Software (if any?)

A hurdle that many of you may be thinking about is that you do not have document automation software or, if you do have it, that you have not had success with it.

Scenario 1: You have document automation software

Count yourself lucky!  If you are actively using document automation software, you can skip to the end.  However, if you feel you are not using it to its full potential (or at all) keep reading. 

There are several things you can do to "dust-off" your document automation platform and get it ready for new automation projects:

  • Re-acquaint yourself with the software that you have and re-establish contact with your vendor;
  • Identify how the software has been used in the past and why projects were or were not successful;
  • Determine whether you have anyone internally who knows how the software works and whether you can do the programming in-house or need external assistance;
  • Ask your vendor for help with upgrading the software as needed and providing training to internal individuals who will be tasked with your new automation projects.

Scenario 2: You do not have document automation software

Do not despair!  There are a number of ways you can still push your automation projects forward.  Three possibilities include:

Acquiring your own software:  Licensing your own subscription to a document automation platform may or may not be feasible. Like any acquisition of legal technology, it is a project itself and is not likely to be a snap decision. 

Looking to other parts of the business:  If the legal department does not have a document automation platform, another department might.  Document automation software does not need to be designed for legal in order to work for many legal documents.  Reach out to your colleagues and see if they have an automation tool you could leverage - even as a pilot project.

Looking to external counsel:  Many law firms will have "client licenses" for their document automation software.  This means that we can work with clients to automate their documents using our software which provides several advantages.  Clients do not have to make significant investments in a platform themselves, they can engage us to leverage our expertise in automation and they only license what they need. 

You Can Have Success

Document automation is a legal technology that requires a plan, some expertise and much perseverance.  But when you identify the right use case, work with lawyers that have the right legal expertise, and have a group of stakeholders with a vested interest in the project being successful, automation and all of its benefits can be achieved.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.