Richard Susskind's question remains useful if it leads to a careful examination of legal work. Which tasks can be performed differently, what happens to quality and responsibility, and who benefits? A book's forecast and a model's benchmark result are different kinds of evidence.
What the argument says
The publisher's description of the revised 2010 edition of The End of Lawyers? presents a forecast that technology and commoditisation will change legal services, substantially erode and often eliminate the jobs of many traditional lawyers, and create new roles. It would be inaccurate to reduce that argument either to the disappearance of every lawyer or to a claim that no jobs would be lost.
The description and contents are a basis for identifying the book's themes. They are not a current census of legal employment or evidence that a particular Swiss firm's services can be automated.
What later research can test
Remus and Levy's 2015 study combined analysis of software capabilities with time-allocation data from large law firms. It challenged sweeping replacement forecasts and emphasised differences between tasks. Its technology assumptions predate today's generative systems, so its estimates should not be reused as a 2026 employment forecast.
More recent evidence concerns bounded tasks. A 2026 randomised study by Schwarcz and colleagues found quality and productivity gains with specified research and reasoning tools among upper-level law students. It does not measure Swiss firm profitability, long-term skill development or the number of lawyers an economy will need.
These findings support a narrower conclusion: some tasks can change substantially, while effects on an entire service and its workforce require further evidence. Faster drafting alone does not tell us whether demand, fees, review work or employment will rise or fall.
Evaluate the whole matter
For a repeatable service, map intake, fact gathering, research, drafting, review, communication and follow-up. Identify where a tool changes the work and where it creates new checking or coordination. Include failed outputs and missed qualifications in the comparison.
Faster first draft
What it may show: Less time in one step
What it does not establish: Lower total cost or unchanged legal quality
Good benchmark result
What it may show: Performance on specified tasks
What it does not establish: Fitness for every jurisdiction or client
More work per person
What it may show: Capacity in the tested workflow
What it does not establish: Future demand or a staffing decision
New service proposal
What it may show: A possible delivery model
What it does not establish: Client willingness to buy or recurring revenue
The table is Ada Studio's analytical aid, not a measured classification of the Swiss legal market. Boutique size, a specialised practice or an expensive subscription does not by itself establish an advantage.
Keep responsibility and learning visible
The Swiss Bar Association's February 2025 AI guidance (in German, French and Italian) addresses supplier arrangements, secrecy and critical review. It does not certify a Swiss-hosted vendor as compliant. Assess the actual service and matter before introducing client material.
If a tool changes junior work, evaluate how people learn to identify missing facts, challenge sources and exercise judgment. Do not assume that removing a task automatically produces equivalent strategic work or preserves its training value.
Questions for a practice review
0/5Key Takeaway
Get in touch to discuss a review of your firm's workflows.