Facts are friends
When legal tech evangelists accidentally reveal the endgame
I have just read the interviews with Max Junestrand, CEO of legal tech firm Legora that Maria Ward-Brennan posted on CityAM (links in sources, below).
While ostensibly promoting his AI platform for law firms, I wonder if he’s inadvertently revealed something rather more profound about where this is all heading.
The telling quote
Junestrand made this observation about future legal hiring:
“Knowing the law, yes, that will be one thing, but AI will also know the law. So the question is, what do you bring on top of that?“
On the surface, this sounds like the usual tech-sector optimism about human-AI collaboration. But read between the lines, and you’ll spot something more unsettling.
If AI “knows the law“ just as well as lawyers do, and if the differentiator becomes “what you bring on top of that“, aren’t we essentially admitting that core legal knowledge - the foundation of our profession - has become commoditised?
The job creation illusion
Linklaters has hired 20 specialist AI lawyers to support a global rollout.
Junestrand is understandably excited about these “legal engineers“ – ex-practitioners who help implement AI systems. He enthusiastically endorses these as exciting “new paths for lawyers that never existed before”.
He praises the firm for “internal ways of working“ - in other words he congratulates the humans for adapting to accommodate the machines.
But two uncomfortable truths underlie his comments.
First, these jobs are designed to replace the work lawyers currently do.
Secondly the “legal engineers” have been recruited to do themselves out of a job.
I don’t believe anyone will disagree with the first of these truths. AI is implemented to do that work that lawyers used to do more efficiently.
The second truth reflects that the engineers are being paid to build workflows and enhance the AI until eventually it has learned enough from them and become sufficiently sophisticated for it (the AI) to build its own workflows. That means that most (if not all) of these roles are transitional (and therefore temporary).
Before discounting this assertion, reflect on the reports that in May Microsoft announced the layoff of about 6,000 employees, a majority of whom were reported to be programmers and put that in the context of the previous month’s comments by CEO Satya Nadella that around 30% of the company’s code was then being written by AI. More recently OpenAI reports that its latest model now successfully replicates 42% of its own coding tasks. The trend is clear. This month Anthropic published its internal survey of engineers including one reporting :
“It kind of feels like I’m coming to work every day to put myself out of a job”.
The Robin AI lesson
Junestrand attributes Robin AI’s recent troubles to their running “two parallel business models“ - direct legal services alongside software provision. The implication is pick one or the other: either you’re a law firm, or you’re a tech company.
This binary vision does not allow room for hybrid models in the long term. You’re either replaced by the technology, or you’re selling it to others (who will be replaced by it). Take your choice.
The illusion of value and irreplaceability
As lawyers we might comfort ourselves by saying that what makes human legal judgement genuinely irreplaceable isn’t efficiency or document processing speed, but the socio-psychological skills and semantic dexterity that allow lawyers to navigate ambiguity, exercise proportional judgement and provide counsel that balances legal technicality with commercial reality. As a lawyer, I enjoyed writing that - as a reader it may make you wince. That’s the problem in a nutshell: we mistake our ability to articulate complexity for evidence that the work itself is irreplaceable.
What I am highlighting is that these value skills are fundamentally interpersonal and cognitive capabilities, they are not legal ones. In fact, while some lawyers - the traditionally highly prized rain makers – may have these skills, they are not attributes typically associated with the legal specialist who has the deep understanding of highly technical issues arising from the law. Those specialists are right in the firing line.
In addition to the “soft skills”, deep domain knowledge is valuable. Understanding the commercial reality of the parties allows advice to be given in context. AI might know contract law perfectly, but if it doesn’t know the particular relationships it will not be able to take account of them.
And that is where Junestrand may have really let the cat out of the bag: if engines like his mean that AI has access to all available information - client histories, industry contexts, relationship dynamics, past disputes - it will eventually be able to emulate those supposedly human skills, even if it doesn’t actually possess them.
The highly prized domain experience is just pattern recognition applied to commercial contexts. And pattern recognition is precisely what AI excels at.
Summit Shah, CEO of Indian e-commerce firm Dukaan, demonstrated this perfectly when his data scientist built an AI chatbot in two days that immediately replaced 90% of the company’s customer support staff. The domain expertise those workers had accumulated over years was reduced to patterns the AI could replicate instantly.
The uncomfortable endgame
In his interviews Junestrand may have revealed more than he intended.
His vision of AI that “knows the law“ combined with his emphasis on implementation and workflow management suggests a future where:
Legal knowledge becomes a commodity available to anyone with internet access
Lawyers become expensive middleware between clients and AI systems
The profession fragments into a small elite who own the technology and everyone else who serves it
Just how long do we have?
For me the most unsettling aspects of this technological development are the increasing scope and accelerating speed of trajectory.
At the current pace of improvement, projections suggest AI could match human experts across the board by 2027 (perhaps before today’s new training contracts are even completed).
But regulatory friction and institutional inertia offer lawyers (some) respite. Legal practice is heavily regulated and regulatory frameworks move glacially compared to technological change.
The entire apparatus of professional oversight was designed for a manual era. Until these frameworks adapt - and that may take years, possibly decades - the conventional lawyer-client relationship will remain the legally recognised model.
There’s also profound institutional conservatism in high-stakes transactions. When a client perceives high value (billions at stake in an M&A deal or the house sale which represents an individual’s highest ever value transaction) clients are remarkably reluctant to experiment with novel approaches. They default to the reassurance of established precedent. “We’ve always done it this way“ becomes a powerful brake on innovation.
The combination of regulatory frameworks and human conservatism buys the profession time - respite, but no cure.
Who will win, ultimately?
Ultimately lawyers, consultants and developers will need to confront an uncomfortable truth: law is fundamentally about risk allocation and mitigation.
Clients don’t really hire lawyers primarily for their knowledge - they hire us because we carry professional indemnity insurance. We are, in essence, an insurance product wrapped in expertise.
This is why the final gatekeepers in this transformation won’t be law firms, clients or even regulators - it will be the insurance industry.
When insurers become comfortable underwriting AI-generated legal advice at comparable premiums to human advice, the economic case for human lawyers largely evaporates.
The profession’s future timeline isn’t determined by when AI becomes technically capable, but when insurers decide the risk is acceptable.
By accepting the risk, insurers introduced cyber insurance 20 years ago, and in doing so they essentially accelerated expansion of the entire digital economy by allowing businesses to transfer cyber risk rather than avoid digital transformation entirely. More recently, drone insurance enabled the commercial UAV market to grow 125% in just three years by accepting risks that many thought uninsurable. That insurance didn’t just grow alongside the drone market - it enabled it.
Once insurers were willing to underwrite AI-generated legal advice, the entire economic case for human lawyers could shift almost overnight.
The change for law could be a decade away, or it could be much sooner than we think, particularly if AI platforms begin offering their own indemnity insurance as part of the package. At that point, Junestrand’s vision becomes reality: lawyers become middleware, and then disappear entirely.
Conclusion
If all goes to plan - from the tech sector’s perspective, not ours - the “lawyers” who survive will be those who own the platforms, not those who use them. Everyone else becomes a glorified prompt engineer, translating client needs into AI instructions, until they too become redundant.
I believe we have established that the question is not whether our profession will be transformed, but whether it will survive in any recognisable form.
Insurers move faster than regulators and care less about precedent than clients. When they’re ready, the profession’s economic foundations will shift overnight. I suspect we’ll find out sooner than we might like.
What’s your view? Has the legal tech sector accidentally revealed its true intentions?
Sources:
https://www.cityam.com/legora-ceo-law-degree-is-not-enough-ai-knows-the-law-what-do-you-bring/
https://www.cityam.com/legora-ceo-troubled-robin-ai-had-two-parallel-business-models/
https://tech.co/news/companies-replace-workers-with-ai
https://decrypt.co/341482/ai-isnt-taking-job-yet-might-soon-openai-data-suggests
https://www.entrepreneur.com/business-news/anthropic-study-heres-how-ai-is-impacting-work/500407
https://www.cnn.com/2023/07/12/business/dukaan-ceo-layoffs-ai-chatbot/index.html
https://www.alliedmarketresearch.com/drone-insurance-market-A323694
https://datahorizzonresearch.com/droneuav-insurance-market-43588