Reflections from LegalTechTalk TechTorget’s side event
We spent last week in London at the TechTorget Nordic side event, held before LegalTechTalk. It brought together leaders building and leading change in the legal sector, and the discussion turned out to be one of the most useful we’ve had on AI and legal work.
A big thank you to the Team TechTorget, Dorte Carlsson, Erik Ødegaard, Helena Hallgarn, and Peter van Dam, for an excellent event and for creating such a valuable discussion around the future of legal work.
It was inspiring to hear so many perspectives on how AI is already changing the legal sector. Our main takeaway was how the legal AI discussion is moving from tools to organisational change.
We have moved past the key question of what AI can do. The main discussions were around how legal work changes, how legal organisations should be managed, and ultimately how value is created when AI becomes part of everyday legal delivery.
Four themes especially resonated with us, because they align closely with what we at Ilves are building for our clients.
AI activity is not the same as AI strategy
Many organisations are experimenting with AI tools, pilots, workshops, prompt training and use case mapping. This is important, but it is not yet a strategy.
Strategy is about what an organisation wants to change and become to build their competitive edge. It begins with understanding how work happens today.
The real questions are: How do busy lawyers change their working habits and thinking? Where does AI genuinely improve quality, speed and profitability? Where are new roles, processes and management data needed for both humans and AI?
A tool can be bought. A new way of working has to be led. In the AI era, strategy means intentionally redesigning the organisation around how work should be done.
You can’t lead change you can’t see
If we do not know what work is being done, who is doing it, in what context, with which tools and at what cost, it becomes very difficult to lead change.
This matters even more when parts of legal work become faster, automated or AI-assisted. Old metrics are no longer enough. Old-fashioned manually entered time entries, without activity data attached, tell only part of the story. Law firms need a richer view of the reality of work. At Ilves, we refer to this capability as Legal Work Intelligence.
For AI adoption to work, a firm needs its own work data, context, metadata and lawyer validation.
The billable hour discussion is a management model discussion
The billable hour discussion always returns but the most important question is not whether the billable hour will disappear.
The deeper issue is that the hour has become much more than a billing unit. It is the operating system for targets, bonuses, profitability and career progression. AI challenges that logic. If AI makes parts of the work faster, how should the service be priced? If delivery is increasingly divided between lawyers and AI tools, how do we understand the true cost of production and where value is created?
As firms move towards fixed fees, value-based pricing or productised services, they need a clearer understanding of the components of legal work, how those components create value, and what it costs to deliver them.
Time may remain part of billing but in the AI era, legal work needs a richer management model and better data to support it. You cannot price outcomes if you do not understand your own cost of production.
Other industries can tell you what each part of a process costs and how long it takes. Law mostly cannot, because it has relied on people manually logging their own hours after the work, which tells you very little about how the work was done. Before a firm can move to fixed fees or value-based pricing with any confidence, it needs to see its own work clearly first.
AI also introduces a cost that traditional timesheets were never built to capture. When part of the work is done by a lawyer and part by an AI agent, hours alone no longer explain what a piece of work cost or where the margin went. Understanding who did what, with which tools, in what context, becomes part of understanding profitability itself.
The billable hour’s role is changing. As one alternative fee arrangement among several, it is not going anywhere, it is evolving into something better. Invoices based on hours typed in from memory are not transparent. The better version is having invoices based on real activity data that show what was done, including the work done with AI.
AI forces us to define the value of lawyers more clearly
One of the most strategic questions discussed was what part of legal work is production, and what part is judgement.
We also explored a related angle at the Legal Design Summit Brainfactory side event on Monday 15 June with Mia Ihamuotila and Malin Männikkö: what types of thinking and skills belong to judgement work, and what belongs to production work.
AI can support, accelerate and partly automate many production-oriented parts of legal work but it does not remove the lawyer’s accountability, responsibility or liability. It makes the roles and responsibilities of both humans and AI systems more important to define.
This also affects junior lawyer development and career paths. As traditional learning has been built around production work, firms need to ask how future lawyers will systematically learn judgement earlier — while also developing talent that can build and maintain the new legal production system.
As production becomes increasingly automated, professional judgement and learning become increasingly dependent on access to the right context, knowledge and organisational experience.
What is next for the legal industry?
TechTorget reinforced our view that the next phase of legal AI will not be solved by one tool, one model or one pilot.
How well legal organisations understand their own work and whether they are willing to build a new layer of management infrastructure around that understanding becomes the important topics to address.
This is the problem Ilves is focused on solving.
We believe that the law firm of the future will need a legal work intelligence layer: a data-driven foundation that helps firms continuously learn from their work, improve professional judgement, and continuously improve how value is created for clients.
Thank you again to the organisers and everyone who contributed to the discussion. This is a conversation well worth continuing.
For more insights follow our CEO Heikki Ilvessalo on LinkedIn.
