Transactional work is weird.
The expertise lies at the top and middle, money is made by work at the bottom.
If you are a senior corporate lawyer, you make money commensurate to your wild success only if you have an army of minions billing time doing more inane mind numbing things.
Making notes, preparing first drafts, proofreading, dressing up for meetings, being available, nodding, smiling, travelling, briefing other lawyers, conferences, retreats, all manner of coordinating things.
Not the sort of thing that moves the needle in the high stakes games that weighs on the experienced shoulders of the partners (or now, senior partner), that reflects actual value creation in the business. This kind of disproportionate value creation makes clients tolerate high hourly rates charged out for minions things like - fixing grammar and aligning paragraphs.
This is not to say that nicely formatted documents with English done right, no grammatical errors and a drop of elegance have no value.
They do.
Nice looking drafts fool readers into overlooking substantive flaws.
A document with paragraphs not aligned, headings all over the place reeks of a disorganised mind without attention to detail. It is another matter that the draft may be legally sound and stand up in court for what it actually says.
There is also another thing.
Automating this used to be terribly difficult.
Until at least mid 2020, there were not so many reliable ways for a computer to understand where one sentence begins, another paragraph ends, or what the structure of a document is or should be.
It was frustrating for those like me working to build solutions to review documents. No matter what we did, we would hit a dead-end, called the segmentation problem.
The segmentation problem in short is - dividing up text in logical ways and giving them headings.
Photo by Jonas Stolle on Unsplash
College graduates performed much better, which worked out well for law firms - an immediately monetizable skill that is also done reliably.
Something to bill the minions out for, to pad the halls of high legal expertise, apart from research, note taking and listening in on calls.
Why not?
Things are finally changing, for the better
This year’s language models are doing better at cracking the segmentation problem.
We might not be far away from having a language model that excels at figuring out where paragraphs should start and end, what headings to give them, and applying standard font styles to documents.
This is good news for young lawyers and the legal profession in general.
Don’t grieve that the ground is shifting under your feet.
Formatting documents is a waste of a legal education and shifts time away from growing in more productive ways. Such as learning about how to defend startup founders before they get into trouble, or how to use AI to prepare a legal notice.
That I along with many of you have had to do it for decades and still do sometimes, will never be worth remembering.