We are witnessing a sharp increase in talk of artificial intelligence (AI) in many industries, including the legal profession.
But will tools like ChatGPT revolutionise the Will-writing process, making it more efficient and accessible?
We think the answer is yes – but probably not for the reasons you’d think.
We asked an AI tool to summarise how AI can impact the Will-writing process. This is the summary part of what it said:
“AI has the potential to benefit customers in will writing by improving efficiency and reducing costs. However, there are concerns regarding the lack of personalization, data security, and legal liability.”
We’d add – the lack of regulation sets a low bar to beat (you don’t just want a Will that is better than a ‘bad’ Will). And it is far from clear how AI could improve upon the “best judgement” legal precedent documents that are based on case-law.
Or to put it another way: given a choice between (i) Will drafting that comes from The STEP handbook for Advisers and tailored to your situation, or (ii) a Will that is AI generated … which would you choose for you entire wealth and all your wishes?
What AI can do is shine a spotlight on the basic (in)competence of lawyers in dealing with simple advice and process. There have already been a handful of studies in the US that show AI can deliver similar results when faced with basic legal questions as compared to answers from junior law professionals. It is only a matter of time before someone takes a load of Wills, asks customers to re-confirm the 15-20 inputs that most Wills are written off, and see whether AI can do a better job.
So a measure of AI’s effectiveness will the portion of Wills it beats, and with most UK Will-writers having no formal qualifications in Will-writing, we suspect it will show a large portion of the current market as lacking basic competence. Indeed, our Will-review process has identified major issues in around 30% of professionally written Wills, and with a total of around 50% having some form of issues with them)
However, in the short term the usefulness of AI beyond exposing the ‘bad’ side of Will-writing is pretty limited. The first flaw is the assumption that a Will can be distilled to 15-20 basic questions, when our process involves well over 200 questions to fully capture a customer’s situation and wishes. Of course AI is the perfect tool for large dataset problems so increasing the number if inputs is certainly not the limit – the limit is extracting those same bits of data from people. And that brings us to…
Steve Jobs famously said “people don’t know what they want until you show it to them”. This is the skill of an adviser – to take an emotionally sensitive subject (like Will-writing, or financial advice, or taking out a mortgage) and translate it into words that customers resonate with.
So for example, we’d say:
- For Wills in particular, your upfront cost is often the worst metric for value
- It is rare that a quick&easy Will covers your whole situation
- Simple Wills are rarely the best tool to protect your Children’s inheritance
It is the adviser’s skill to help customers navigate the range of decisions and options facing them. Maybe AI will get there one day – but it is certainly a while off.
Putting customer benefit first
But what is the point of a Will anyway? It is to form the legal backbone of the process on death. If there is an area that AI can have a much more immediate impact it is to fix what remains an essentially 19th-Century based manual process of sending letters, summarising assets&liabilities, and calculating distributions.
There are approximately 6,000 banks plus 6,000 other financial institutions in Europe alone… but if your assets are not held by the ~40 or so firms that “professional probate” firms typically contacted, there is a good chance your money will not go to your family.
Could AI scan your email archive after death to identify potentially orphaned bank accounts and policies? Perhaps with some ease, but this relies on access to the email archive of a recently deceased person so is unlikely to happen without law changes (which themselves are unlikely to happen due to privacy concerns). But this reminds us of the #1 rule of Will-writing – to have one. Something is better than nothing. A list of assets (say, stored in your Kinvault) is better than no list of assets, and hoping future-AI will somehow fix your poor record keeping is a fools game.
The simple reality is that it isn’t AI that we need in Will-writing, but I for Intelligence. With under 5% of Will-writers holding Will-writing qualifications, a starting point is to make sure your writer is STEP qualified (STEP = the gold-standard qualification provider in the UK). And then it is about going ‘beyond the Will’, and considering the full range of what your family need you to do to make sure things don’t go wrong on death.
Image credit: MidJourney (AI)