Smarter Documents

June 22, 2016

This excellent event, hosted by Allen & Overy (“A&O”) and chaired by Andrew Dey, looked at the evolution of legal document creation and thoughts for the future. A very good mix of theory and good practice – a hallmark of the SCL KM events.

Document creation today

Nigel Rea from LexisNexis began the event by setting out very clearly the evolution of document creation from free hand drafting, through flat precedents, automated precedents to assisted drafting with embedded guidance and checking.

He offered some very helpful tips on getting started with document automation:

– Keep it simple: LexisNexis have noticed that documents with 150 questions are used less frequently than those with just 10 questions.

– Consider carefully how the document will be used and who will be producing it.  This will inform what sections you automate.

– Don’t be tempted to automate to 100%. Leave the lawyers something to do!

– Start small, focusing on one practice area and include process improvement.

– Capture the data: you can analyse the answer files to see how people are answering the automation questions. You can capture amounts, currencies, common answers. 

– Think about the infrastructure: who will automate, who will support and update?

– Remember to assess how you will measure success and determine the ROI. Document automation exercises are expensive in terms of time.

Nigel also offered some insight into the chronology of the automation process.  He identified six phases:

– Discovery
– Plan and prioritise
– Technology investment
– Automation
– Maintenance and support
– Benefit and continuous improvement

Nigel acknowledged that in reality, the order is often altered and runs thus: technology investment (with no planning), discovery (with no strategy), planning and prioritising (after the event). The result is no buy-in, because there is no clarity around strategy.  This leads to an elongated delivery cycle, which impacts on the profile of the project. Finally, it is hard to measure the benefit because no data is collected.

Nigel then touched on LexisNexis’ routemap.  They have recently purchased Lex Machina, a legal analytics tool originally focused on IP litigation in the States. LexisNexis are focusing on “data driven insight”: all legal knowledge sits in a huge data set, exploiting this data in new and innovative ways is a key focus. Nigel foresees the day when you will be able to select the emotion for a negotiation (hard or soft) and clauses which fit the “mood” will auto select. 

Case studies and best practice

Ruth Ward then shared the A&O document automation journey. Her first role at the firm in 1999 was to promote document automation. It was easy as people were happy to have 10 hours less of drafting to do! It works well for finance documents, but doesn’t always work for other areas. Simple documents and suites of documents are best.

She advised analysing how the automation will add value before you start the process. Planning up front is really important. She suggested asking the following questions:

– What problem are you trying to solve?
– How much will it cost?
– How long does it currently take to produce a first draft?
– How often is the document used?
– Can you improve content management by having one automation for a series of documents?
– Will embedded guidance reduce the amount of queries?
– Will speed and efficiency increase margins?
– Will it reduce risk, with fewer margins for error? 
– Will clients value getting the document quicker? 
– Can you make a suite?
– Do you have the data for all the questions: limit your questions to the absolute essentials – the fewer the better

Lucy Sherwood explained Pinsent Masons’ approach. Automation is just part of the way they work. It is core to the strategy of the business to deliver value. They integrate automation into how they do business and they make a workflow to wrap around the whole process.  It is driven by value and strategy.

It also enables them to work collaboratively with clients.  There is a huge amount of information that could be useful to management or clients hidden in data, for example: how many times is this question answered in a particular way?  This information could be extracted by analysing the information with the use of artificial intelligence. Often it is information that the client would love to have.

She has noticed an evolution: lawyers used to be worried about documentation automation, now they are worried about artificial intelligence!

During the subsequent Q&A a number of issues were highlighted:

– How do juniors learn if they never draft documents freehand? More training was advocated
– How to get by-in: Listen to your sceptics, they will have good ideas  and launch it quickly so people see the benefit
– Who does the coding: central team, consultants, local Practice Groups?
– Clauses v documents? We discussed whether you could have a suite of clauses rather than multiple documents with many common clauses.  Technology allows you to share clauses across documents.  This works at practice level or where you have genuine boilerplate, but that is hard to agree. It should be approached with caution: even in boilerplate clauses the context is really important.

Smarter documents – the future

Andy Wishart, from Contract Express at Thomson Reuters, then looked at the future of smarter documents.  He focused on automation, collaboration and analytics. 

Automation: to speed up document assembly, they are combining automation with other products: PLC content and firm’s house style.

They will also enable firms to combine PLC precedents with their own content in a document automation tool.

Collaboration with clients: he envisages collaborating with clients through the entire life cycle of a matter through a matter dashboard. Once the matter type is identified, it will push a workflow to the dashboard, prompting appropriate documents to be drafted and reviewed, and eventually signed via electronic signature (e.g. DocuSign). Currently red lining is the enemy of this end to end process as it takes the document out of the workflow.

Analytics: if you have a workflow, you have rich data to analyse. The client can move from basic aspects, such as what has been signed and what hasn’t, to more high level insights: deviations from standard wording to better manage risk, finally to a more macro focus: for example trends in their business.

The panel agreed that as firms (and their clients) want to analyse document content more than ever before, creating “smarter” documents is an area of technology that is certain to significantly impact the legal industry in the next few years. Using artificial intelligence to analyse the data created will enable firms to predict outcomes and trends. An excellent evening, with insightful speakers and a good discussion. Those present left with much food for thought in this key area for KM in the legal profession.

Lucy Dillon
Andrew Dey
Shelagh Taylor