AI slop… It's all over your LinkedIn feed, planning that trip in your group chat, and now it's in the legal world too. Even now, as I write this on the train from Liverpool to London, I wonder… will readers think this is AI? After all, it's being published in a legal AI company's newsletter. Anyway… I digress.
I often wonder what implications it will have on the legal world. Some are visible today: companies all over LinkedIn and in their newsletters publicising how effective and efficient AI has been, and how it has supercharged their work. I get it. When I sit down with lawyers, be it friends, colleagues or people at events, they're all excited by the idea, and most can point to something: emails drafted faster, inboxes triaged, documents summarised. But ask them where the time they've saved has gone, and the answer gets vague. They can name tasks AI has made easier. What they struggle to name is a work day that actually feels different.
I recently started a new in-house role at one of the UK's largest private healthcare providers. Before that, I was a commercial adviser whilst building in the tech space, self-taught with Claude, a Railway account and a PocketBase subscription. AI was a huge part of my day. In my new role, until recently, the only AI on our systems was Copilot. That's right, I've been doing law the old-school way: reviewing contracts myself, cross-checking, cross-referencing and advising without AI, albeit Copilot may have proofread a few of my emails.
Now we're starting to look at what's out there, and I'm tasked with bringing AI into the legal team. The reality is the market isn't short of providers, it's short of difference. Most are offering the same thing, packaged slightly differently. I have to admit, I have yet to see a provider I truly think will revolutionise a lawyer's day-to-day. Yes, there is progress. In full disclosure, there are tools I have yet to see in action, Flank included. But what I mostly see is a crop of Claude Code wrappers. That isn't necessarily a bad thing; you can provide huge value with a wrapper if it's executed well.
The Bermuda Triangle
The problem isn't these companies, it's that the legal industry hasn't evolved much in decades. Legal teams still live in the Outlook/Word/shared drive Bermuda Triangle. I use that reference because lawyers fly into these three platforms and never come out. Habits set, and they're very hard to change.
So most AI companies aren't really selling software, they're trying to rebuild an ecosystem lawyers have comfortably accepted for years. Every company is competing at every level: the contract reviewer, the drafter, the CRM, the Outlook. Some integrate with those tools; some try to replace them.
Here's the thing though. The triangle isn't going anywhere, and it isn't only lawyers in it. The business lives there too. Sales forward a redline by email. Procurement attaches a Word document. The signed copy ends up in a folder somewhere. Any tool that asks people to log in somewhere new isn't just asking the legal team to change their habits, it's asking the whole company to.
The crossroads
Legal AI is at a crossroads, and I see two roads.
The first is the inevitable BlackBerry-to-iPhone moment: a product that actually works out of the box. No "forward deployed engineer" required to make it usable. (It's all the rage now. The running joke is that fractional lawyers are being rebranded as "forward deployed lawyers".) The iPhone didn't ship with an engineer. It was something not just the tech-savvy could use, and legal AI needs the same: a tool partners can pick up too. Yes, even the partner who still can't convert a document to PDF.
The second is the one I'd bet on: stop building tools for lawyers to use, and start building agents that work inside the triangle on legal's behalf. An agent that picks up the email, reads the attachment, chases the missing contract value, produces the first draft or first-pass redline, and files it where it belongs. The lawyer stops being the first pair of hands on routine work and becomes the person who signs off the exceptions.
That matters because lawyers usually see a contract a long way into its life. Procurement, sales and the wider business own most of it before it reaches legal, and they're contract-smart people with almost nothing built for them. An agent in the inbox they already use meets them at the start, not the end. It's also why this road will be niche by nature. A procurement-heavy healthcare group, a property litigation team and a marketing agency don't need the same agent.
But an agent is only as good as the positions it's been given. It can apply a playbook. It can't write one. What a company will and won't accept, what it refused last time and why, still lives in people's heads and old email threads. No deployment timeline covers that bit.
Where did the time go?
The big players will win the large corporates. That's where the budgets are, and where the iPhone moment will most likely land first. But if AI is going to change our field the way email replaced the fax machine, it can't stop there. It has to work for the in-house team of three as much as the large firm, and it has to change the work day of every lawyer, not just the ones with an innovation budget.
That, for me, is the real test. Not how many features a platform ships, but whether the lawyer using it can answer the question my friends and colleagues can't… where did the time go? The honest answer should be: into the decisions only a lawyer can make. When that's the answer, we'll have our iPhone moment, and the question won't be whether AI wrote this, but whether it was right.
Ahmed is an in-house commercial solicitor at a UK private healthcare group. He handles procurement, technology and commercial contracts, and leads his legal team's work on bringing in AI. He previously advised startups and founders as a commercial adviser, and is speaking on legal AI at the Kiss the Future AI Summit in Sarajevo this October.
