Hi Damien! A pleasant read as always. I really found the idea of insurance for risk caused by AI systems interesting. This is because mistakes caused by such systems can be costly in service sectors such as law, medicine, consulting, and personal financial advice. Furthermore, with respect to the billable hour, I think this is one of the reasons why ALSPs will adopt AI more widely, because their workflows are already standardized and have a fixed-fee structure. The competition for market share has never been higher, I feel, with so many entrants, including AI-native law firms, BigLaw, ALSPs, Mid-sized law firms, and Specialized boutique verticals.
You are 100% right that this is giving a new impetus to new entrants, at least in markets where there can be new entrants. Next question is whether legal services will continue to be liberalised, or whether we'll see a pushback.
I feel there will be pushback, given how the billable model is closely tied to lawyers' identities. The argument I would put forth to alleviate these concerns is that a fixed-fee structure supports longevity in client acquisition and increases the number of matters/transactions the firm closes (Higher revenue), depending on the firm's type. Furthermore, with any technological change, incumbents generally take longer to adapt to new technologies and business models. Newer entrants such as Soxton AI and Moritz are charging fixed fees based on contracts reviewed and drafted. Time will tell but this is just my opinion.
Hi Damien, strong synthesis and great piece as always.
AI note-takers, synthetic grievances and the economics of adoption look like separate strands but share the same underlying problem. AI is reducing the cost of producing records, complaints, outputs and process artefacts faster than institutions can adapt their verification systems.
On note-takers, I agree the privilege panic can be overstated. The meeting platform is already processing the audio and video. The harder question is not whether a record exists but what kind of record it is. An AI transcript is automatic, persistent and decontextualised. It is often treated as more authoritative than it deserves to be. That is where discovery risk and hallucination risk start to overlap.
The synthetic-grievance point is even sharper. If a complaint can be generated at near-zero cost but must still be investigated through a human professional process, the asymmetry is obvious. The bottleneck moves from production to verification. Insurance entering the market is a clear signal this is no longer theoretical.
The Shaver point is the one I think matters most for law firms. Culture is part of the story but it is not the structural problem. The billable hour is not just a pricing mechanism; it is the internal value-allocation system. Until firms work out how to redistribute the gains from AI investment, adoption will remain performative at the edges and contested at the centre.
The profession keeps talking about AI as a technology problem. This piece shows why it is increasingly an institutional-design problem.
Hi Damien! A pleasant read as always. I really found the idea of insurance for risk caused by AI systems interesting. This is because mistakes caused by such systems can be costly in service sectors such as law, medicine, consulting, and personal financial advice. Furthermore, with respect to the billable hour, I think this is one of the reasons why ALSPs will adopt AI more widely, because their workflows are already standardized and have a fixed-fee structure. The competition for market share has never been higher, I feel, with so many entrants, including AI-native law firms, BigLaw, ALSPs, Mid-sized law firms, and Specialized boutique verticals.
You are 100% right that this is giving a new impetus to new entrants, at least in markets where there can be new entrants. Next question is whether legal services will continue to be liberalised, or whether we'll see a pushback.
I feel there will be pushback, given how the billable model is closely tied to lawyers' identities. The argument I would put forth to alleviate these concerns is that a fixed-fee structure supports longevity in client acquisition and increases the number of matters/transactions the firm closes (Higher revenue), depending on the firm's type. Furthermore, with any technological change, incumbents generally take longer to adapt to new technologies and business models. Newer entrants such as Soxton AI and Moritz are charging fixed fees based on contracts reviewed and drafted. Time will tell but this is just my opinion.
Just stumbled across your stack last night. Interesting stuff and a great resource. I mentioned a bit of your data in my recent post. Cheers.
Thanks !
The idea that some things should neither be written down nor recorded is fundamental to survival.
It’s good that you point this out to those who lose track of the basics, when caught up in the allure of the technologies.
Hi Damien, strong synthesis and great piece as always.
AI note-takers, synthetic grievances and the economics of adoption look like separate strands but share the same underlying problem. AI is reducing the cost of producing records, complaints, outputs and process artefacts faster than institutions can adapt their verification systems.
On note-takers, I agree the privilege panic can be overstated. The meeting platform is already processing the audio and video. The harder question is not whether a record exists but what kind of record it is. An AI transcript is automatic, persistent and decontextualised. It is often treated as more authoritative than it deserves to be. That is where discovery risk and hallucination risk start to overlap.
The synthetic-grievance point is even sharper. If a complaint can be generated at near-zero cost but must still be investigated through a human professional process, the asymmetry is obvious. The bottleneck moves from production to verification. Insurance entering the market is a clear signal this is no longer theoretical.
The Shaver point is the one I think matters most for law firms. Culture is part of the story but it is not the structural problem. The billable hour is not just a pricing mechanism; it is the internal value-allocation system. Until firms work out how to redistribute the gains from AI investment, adoption will remain performative at the edges and contested at the centre.
The profession keeps talking about AI as a technology problem. This piece shows why it is increasingly an institutional-design problem.