A recent development in the area of legal tech has been the rise of technology capable of providing functional tools for lawyers reviewing legal documents. While some of these tools, such as Luminance, are designed to assist lawyers in the review of data, such as contracts[1], others seek to go one step further, by automating the drafting process itself. The technology behind these initiatives is currently being tested by a number of different firms within the profession. This raises a number of questions, from the benefits and limitations of such programs to their applicability to various sectors of the legal field. We must also consider what the role of the future lawyer will be if such technology becomes widespread.

                                                                                                                                     Issues with Traditional Document Drafting

Before considering the potential of automation in this area, it is worth briefly analyzing what traditional legal document drafting looks like. A very large percentage of these documents have already been automated at least to a certain extent. Many are created digitally by relying on precedents and model documents. The process is not only inherently complex, as the drafting of these texts inevitably requires the consideration of intricate legal problems, but it can also very rapidly become a tedious and labor-intensive process. In order to draft a business contract for instance, lawyers might have to sift through hundreds of documents to determine which particular type of clause their new template should be based on. If these tasks are performed manually, there is no system that can automatically indicate to legal professionals which knowledge is most likely to be relevant to their current draft. In light of these issues, what additional benefits do current automation initiatives offer to legal drafters and clients in this area?

CC Dr@ft

One example comes courtesy of Dr@ft, Clifford Chance’s document automation application, which exists as a managed version of Thomson Reuters’ Contract Express product. The basic concept of the software, offered as a service to clients and used within the firm itself, is to remove the need to base future drafts on pre-existing model documents. Instead, the user is presented with an online questionnaire. The answers generate a variety of documents, themselves based on the answers provided, from which the user can choose from. At later stages, when related documents need to be drafted, most of the information provided can be reused[2]. This straightforward process is made possible by the manual coding of conditional spans, which specify a set of conditions that must be met for a particular piece of text to be displayed[3].

While these methods do represent a significant step in the automation of document drafting, they do not yet involve natural language processing. Already widely used for tasks such as contract review, statistical pattern recognition techniques that analyse the language of a particular document and predict how to recognise and apply it in the future are not as of now a part of the drafting process. Regardless of whether AI of this sort could be incorporated by applications such as Dr@ft in the future for the creation of documents, the platform already represents a significant innovation in this area. Furthermore, the addition of decision trees that are triggered by the user’s answers to the questions generated by the software contribute to the production of a tailor-made document template[4].

                                                                                                                                  Benefits of Legal Tech in Document Drafting

There are several benefits to the use of such applications in the drafting of legal documents, beginning with time-saving. Clifford Chance estimates that Dr@ft users can look forward to spending at least 50% less time on drafting tasks[5]. Another important, often overlooked added value of using this solution to draft legal documents is consistency. As previously mentioned, information entered to create one template can be reinserted to produce other related documents[6]. This allows for the creation of one set of documents based on the same initial questionnaire, improving the quality and regularity of the user’s first draft. These benefits combined bring us to a key incentive for widespread adoption of similar tools in the legal industry: the reduction of internal costs. Compliance departments inevitably see their workload decrease, and lawyers see their time freed up for use in other revenue-generating tasks.

Furthermore, certain costs associated with specific sectors of the legal profession can significantly be reduced. Automated drafting of defences and witness statements can replace the need to hire a barrister to perform these tasks. Such advantages help to explain the proliferation of document drafting automation in the legal tech space. Other examples include Avvoka, a platform offering a range of services for contract automation, including options for drafting documents with a particular emphasis on facilitated negotiations with counterparties[7]. Last year, the legal tech startup Genie AI received £2 million in funding for its intelligent contract editor, which hopes to allow lawyers to channel the collective knowledge of their firms and apply the relevant knowledge when drafting documents[8]. There is little doubt that similar initiatives will continue to be sparked as this space continues to develop.

                                                                                                                            Limitations and barriers to widespread adoption

Why then, given all of these exciting benefits, are platforms such as Dr@ft or Avvoka still not the status quo in the legal profession across firms of all different sizes? For starters, there are certain technical barriers to wider automation of legal documents’ drafting. Although the drafting alternatives that we have discussed so far remove the need for lawyers to rely on previous models to draft new documents, they themselves operate on the availability and the quality of standard documentation in a particular area of law. In fields such as employment and capital markets, where there is no shortage of such documentation, the process is made all the more efficient and quick. However, because platforms such as Dr@ft cannot (yet) create their own standard documentation or model templates, they still rely on the data and inputs provided by humans. As a result, the cost of automation is harder to justify when the document in question isn’t used often enough. The first result is that smaller firms, which have less documentation capital at their disposal, are less likely to benefit from the marginal returns of the software and more likely to be deterred by its fixed cost. These shortcomings also create a disequilibrium in incentives for automating documents in certain fields (capital markets, employment) rather than others. The current lack of uniformity in the applicability of automated drafting technology to different sectors of law acts as a barrier to widespread adoption across the profession.

Aside from the fact that there are still tasks document drafting technology cannot achieve, the rise of these alternatives poses a broader question: what will the role of the lawyer be in this world where so many tasks traditionally associated with the legal profession have been so significantly automated? One perspective is to view these tools as “super assistants”, performing tasks at an incredible speed and precision, but still part of a pecking order in which lawyers are at the top and play an active role in reviewing, analysing and acting on the information presented to them. There is certainly truth to this statement. Applications like Dr@ft still require users to choose which document templates they prefer, and are still based on a questionnaire that requires a human input. Nevertheless, this doesn’t address the other disruptive consequence of automation in this space: how will lawyers utilise the 50% or higher proportion of their time that will no longer be spent drafting and reviewing documents? An optimistic view is that this time will be used to focus on the more complex and intricate legal and commercial implications of the work that lawyers engage with. Without discarding this perspective and the realistic view of the future it puts forward, such a dramatic change in the way workers within the legal profession operate nevertheless remains, as of now, a barrier to adoption. For those who wish to surmount it, legal work in the future will require, in addition to a fundamental reevaluation of the way employees in the field spend their time, considering these platforms not simply as assistants but as collaborators, who will eventually be able to learn and make their own conclusions from the data that they are fed.

Author: Raphaël Beaulieu

Edited: Panteleimon Athanasiou

A special thanks to Chris Ireland for his contribution to this article, specifically through his first-hand experience using Dr@ft.

Further Reading:

Ashley, Kevin D. Artificial Intelligence and Legal Analytics: New Tools for Law Practice in the Digital Age. Cambridge University Press, 2019.

Beaulieu, Raphaël. “Luminance: The AI Toolkit of the Modern Lawyer.” Medium, Medium, 25 Mar. 2020, medium.com/@k1922409/luminance-the-ai-toolkit-of-the-modern-lawyer-a25f556dfab2.

Lightfoot, Chrissie. Tomorrow's Naked Lawyer: NewTech, NewHuman, NewLaw: How to Be Successful, 2015 to 2045. Ark Group, 2014.

Susskind, Richard E., and Daniel Susskind. The Future of the Professions How Technology Will Transform the Work of Human Experts. Oxford University Press, 2017.


[1]Artificial Lawyer. “How Slaughter and May Is Making Use of Legal AI Across The Firm.” Artificial Lawyer, 28 Mar. 2019, www.artificiallawyer.com/2019/03/28/how-slaughter-and-may-is-making-use-of-legal-ai-across-the-firm/.

[2]“Artificial Intelligence and the Future for Legal Services .” Clifford Chance , Nov. 2017.





[7]“Homepage.” Avvoka, Avvoka , 11 June 2020, avvoka.com/.

[8]O'Hear, Steve. “Legaltech Startup Genie AI Scores £2M Seed for Its 'Intelligent' Contract Editor.” TechCrunch, TechCrunch, 23 July 2019, techcrunch.com/2019/07/23/genie-ai/.