Generative AI – The Future of Document Automation and Precedents (October 2023)
There has been a huge amount of speculation about Generative AI and the affect it will have on different suppliers. Some have predicted it will mean the death of document automation and precedents.
On the one hand you can see their argument (i.e. why would anyone spend time developing and testing a system when they can simply point Generative AI at their internal documentation and generate a document in a fraction of the time) – the Thomson Reuters Copilot plug in may make this even more the case. In recent times the document automation industry has been well served; we have some very good suppliers who adopt a whole range of models including no code, low code, pro code, managed service etc. Our view is that the impending demise of these systems could not be further from the truth. Document automation and precedents are likely to have a very bright future.
Within a business you are likely to have a whole range of agreements of varying importance and volumes. If there is an agreement you are producing a lot of and the first drafts have to be absolutely correct, you should absolutely use document automation (or case/matter management). This means that the document can be generated quickly, you have certainty about the data points and certainty about the output. Plus, you can link this to other systems such as Salesforce. If the product has been properly tested, you can avoid the need to give the output a thorough page by page review (which of course takes time). By contrast if there is a document you do not produce often (and accordingly you can’t justify automating), Generative AI may have a place (e.g. for example you could focus the Gen AI at precedents (as opposed to drafts used on previous deals). What you have to accept when looking at such an output document though is that you will need to thoroughly review it, page by page and/or would need to verify it against a trusted precedent using a playbook tool or Gen AI itself – in time we will probably see more advanced tools and prompts which show the data lineage – i.e. what amendments were made to what precedent? Without this though the time taken to do a review (depending on the length of the document) could be longer than using a document automation tool but equally may still be worthwhile given that you can’t justify fully automating a document of that type due to the low volumes. Also, will Generative AI ever deal with the advanced conditionality you see in document automation systems even if you charge it with asking you the questions it would need to know to draft a good first draft?
What I am saying therefore is that document automation and Generative AI should co-exist. They are all part of an ecosystem. Precedents instead of being less valuable will be more valuable as a result of an increasing reliance being put upon trusted content. In a world of exponential content generation reliable content will become more valuable.
Our overall advice is to think through what future demand will look like, both from your own business and from clients (it will no doubt shift from what we see currently) and to gear yourselves up appropriately with the right tools. No legal tech tool is a panacea and increasingly we are going to see tools co-existing and working together as a single ecosystem with inbuilt verification systems but people need to be given guidance to ensure they are using the right tools for the right work.
An increasingly important type of document system will be the playbook tool. We have seen these used extensively in-house but less so in law firms. As lawyers will increasingly be asked to check Gen AI documents by clients (i.e. lawyers have always carried out this role as they are seen to have a mastery of the truth and accuracy) it is easy to see how these tools will become more essential and again part of the ecosystem. We are likely to see increased use of Gen AI governance tools too tracking usage and screening in correct uses of data.
Before we conclude, the importance of precedents cannot be forgotten. The legal profession is littered with examples of negligence where a core precedent has not been used as a base but if you listen to some, we run the risk of this happening again if the Gen AI tools are focused on drafts which have been previously used, as opposed to core precedents. First drafts are often focussed on specific deal circumstances and final drafts are negotiated. Everything in between is sub-optimal too as base content for Generative AI. For more on this see The Future of Law Post Generative AI (October 2023) — Hyperscale Group Limited and Generative AI and Avoiding School Boy/Girl Mistakes (October 2023) — Hyperscale Group Limited.
The successful law firm or in-house team of the future needs to develop their strategy in this area carefully. They need to ensure they don’t throw the baby out with the bath water and have a clear strategy for document generation, document automation, playbooks and data hygiene/governance. All will be essential.
For our other recent articles on Generative AI:
Generative AI and Avoiding School Boy/Girl Mistakes (October 2023) — Hyperscale Group Limited
The Future of Law Post Generative AI (October 2023) — Hyperscale Group Limited
Hyperscale Group are an Independent Digital, Innovation, Operational Advisory and Implementation Business with over 25 years plus deep market experience. We work for In-House Legal teams and Professional Services Firms all around the world and support them developing and implementing their strategies. We help our clients to make the right things happen.
Our Experience: Click here - Our Services: Click here - Client Testimonials: Click here
How We Work: Click here
For more information, please contact Dereksouthall@hyperscalegroup.com
