The End of Lawyers? Susskind's Predictions, 15 Years On

Picture this: you run a 12-lawyer firm in Basel. Your practice centres on pharma IP and cross-border commercial work for companies supplying Novartis, Roche, and their suppliers. Your billing rate is CHF 450 per hour. A client calls to say their procurement team has started using an AI tool to draft and red-line standard supply agreements before sending them to you. They want to know if they still need you for that work — and if so, why.

That question is not hypothetical. It is already landing on managing partners across Basel, Zurich, and Frankfurt. Richard Susskind saw it coming when he published The End of Lawyers? in 2008. Fifteen years on, the question mark in his title feels less like rhetorical caution and more like genuine urgency — especially for mid-size firms that the book's BigLaw focus never quite addressed.

What Susskind Actually Argued

Susskind wasn't predicting mass unemployment for lawyers. He was making a structural argument: legal work sits along a spectrum from fully bespoke — complex, novel, high-stakes — to fully commoditised — routine, repetitive, process-driven. His uncomfortable claim was that most of what lawyers actually do falls closer to the commoditised end than the profession admits. Technology, he argued, would steadily move work down that spectrum, while new entrants would capture whatever fell off the bottom.

He foresaw three specific shifts: online legal services displacing high-street solicitors for routine personal matters; large-scale document review being automated; and legal knowledge being packaged into tools that clients could access directly, reducing the lawyer from sole expert to guide or verifier.

The profession largely dismissed this as provocation. Partners at Magic Circle firms were not losing sleep over software. Smaller firms barely registered the argument at all.

What Has Actually Happened

The record is mixed — but the direction is unmistakable, and the pace is accelerating.

Document review and legal research transformed first. By the mid-2010s, e-discovery software demonstrably outperformed junior associate review on both speed and accuracy. Today, Swiss legal-tech vendors such as Omnilex and Swiss-Noxtua are turning legal research into a workflow organised around structured citations, multilingual retrieval, and jurisdiction-specific sources. Noxtua's parent company raised EUR 80.7M in 2025 and launched a Europe License in February 2026. These aren't American imports. They're being built down the road.

The competitive pressure is already measurable. Legora, a Stockholm-based legal AI platform, tripled from 250 firms to over 600 across 50 markets in eight months. It hit a $5.55 billion valuation in March 2026. That rate of adoption tells you something important: the firms signing up aren't Magic Circle firms trying to shave costs. They are firms like yours, discovering that AI tools are now accessible, affordable, and Swiss-compliant enough to deploy.

Goldman Sachs estimated that 44% of legal tasks are automatable — the highest share of any profession except administrative and clerical work. That number matters for your firm in a specific way: 73% of Swiss lawyers work in practices of 1–10 lawyers. The original targets of automation were the associate armies at large firms. But when the work that required 50 associates can be done with AI, the cost pressure shifts — and it shifts toward every firm billing hourly for repeatable process work.

What Susskind Underestimated

Two things, primarily.

First, the pace of institutional resistance. Swiss bar associations, courts, and the SAV's own AI guidelines (adopted June 2024) have created a structured compliance framework, but they've also raised the threshold for adoption. Every AI tool touching client data must qualify as a Hilfsperson under Art. 321 StGB and comply with nFADP. Professional secrecy breaches can create criminal exposure under Swiss law, which is exactly why firms working with providers that have genuinely solved data sovereignty and outsourcing controls start from a materially stronger position.

Second, Susskind couldn't have anticipated large language models. Writing in 2008, he described the direction correctly, but not the velocity. The step-change in the early 2020s wasn't incremental. Pre-GenAI, analysts estimated only around 10% of legal practices were genuinely tech-viable. Post-GenAI, that estimate is closer to 90%. The floor dropped out of what counts as "too small to automate."

What's Coming in the Next Ten Years — for Your Firm

Jordan Furlong, one of the sharpest legal sector analysts writing today, frames the current moment as a bell curve with three populations: a leading edge that's making aggressive moves, a dangerous middle mass running experiments with no clear strategy, and a tail that hasn't looked up yet. The middle mass is the most concerning — not because they're doing nothing, but because pilots without strategy create false comfort.

Richard Tromans, founder of Artificial Lawyer, puts it more bluntly. In the traditional firm model, associates are the production engine — they're the means of scaling partner judgment through billable hours. AI offers an alternative scale mechanism that doesn't grow with headcount. For a 10-20 lawyer firm with no associate army to protect, that's not a threat. It's purely additive.

What this means practically:

Autonomous workflows will handle entire transaction flows for standard deal types — due diligence checklists, contract first drafts, regulatory filing preparation. The lawyer's role shifts to setting parameters, reviewing edge cases, and accepting liability. For a Basel firm doing cross-border pharma supply work, this means your knowledge of Swiss OR obligations and cantonal procedural differences becomes more valuable, not less — because the AI handles the templated layer and you handle the judgment layer.

Continuous compliance monitoring will replace periodic advice for clients with ongoing regulatory exposure. For firms serving Basel's pharma cluster — Novartis, Roche, Syngenta, Lonza — real-time FINMA or Swissmedic monitoring products represent a genuine recurring revenue model that hourly billing can never replicate.

Susskind's fundamental claim was about direction. The legal profession isn't ending. But the shape of legal work — who does what, at what cost, for which clients — is being restructured in exactly the ways he described. The question mark in his title was always the honest part. For managing partners at mid-size Swiss and German firms, the honest answer now is: the question is no longer whether, it's how fast, and whether your firm is on the right side of it.

The Regular Work Band: Where AI Really Bites

A three-tier model of legal work clarifies exactly where AI disrupts: strategic work (~10–15%), regular legal work (~55–60%), and clerical/routine work (~25–30%). The critical insight: AI is not just automating the boring clerical parts — it is automating the entire "regular" middle band that constitutes the majority of billable hours.

And crucially: there is no queue of strategic work waiting. Removing routine work does not automatically generate equivalent strategic demand. Strategic work is driven by client needs, not by lawyer availability. The chess analogy applies: AI can compress the learning curve, but only if firms intentionally design training around strategic capability rather than hoping the work materialises organically.

Mark Smolik confirms that over 50% of corporate law department output is commodity work — NDAs, contracts, leases, standard pleadings. This is where AI-driven automation and alternative business structure competition will hit first. One firm explicitly refused to adopt AI-driven due diligence for Smolik's work — and lost the engagement.

Want to work through these changes for your firm? Get in touch.