From AI-Assisted to AI-Driven
In yesterday’s live session, the legal industry’s rapid shift from scepticism to experimentation with artificial intelligence was front and centre. The more pressing question is no longer whether to use AI, but what it truly means to be AI native.
Moderators Alex Baker, Founder of Legal Tech Collective, and Chris Bridges, Founder of Tacit Legal and Tilda, explored the distinction between firms that are merely AI-assisted and those built AI-first from the ground up.
Bridges defined the difference clearly: an AI-native firm is AI-driven and human-assisted, rather than human-driven and AI-assisted. In practical terms, this means automation and AI perform the majority of the work, with human oversight applied at critical checkpoints. Most traditional firms remain firmly in the latter category, using AI as a productivity layer rather than a structural foundation.
The transition between these models is possible, but difficult. As Bridges noted, attempting to retrofit AI into an existing firm often feels like “building the plane whilst flying it”.
The Structural Challenge for Full-Service Firms
Baker emphasised that many established firms face a structural constraint. Full-service firms operate across multiple practice areas and client types. To become truly AI native, they would effectively need to build multiple start-ups simultaneously, each solving a defined problem for a specific client segment.
By contrast, emerging AI-native firms are highly focused. They identify a narrow, commercially viable problem and build a technology-led solution around it. The result is not simply process automation, but a reimagined service model.
Governance also plays a critical role. Hybrid models can work, but only if AI-driven ventures are run with sufficient independence from traditional partnership structures. Without that separation, legacy governance often pulls innovation back into established operating norms.
Start With Client Experience
Both speakers agreed that client experience is the real catalyst for AI-native thinking.
Baker pointed to Garfield, a UK debt recovery firm serving small businesses, as an example. Instead of following the traditional law firm journey, enquiry, call, engagement letter, document exchange, clients upload documentation directly into a platform, with much of the process automated. Human lawyers remain involved, but the interface is fundamentally digital.
The shift reflects changing client expectations. A generation accustomed to seamless digital services is unlikely to accept unnecessarily complex legal onboarding processes. AI-native firms, therefore, rethink not just internal workflows, but the entire client interaction model.
Supervision, Insurance, and Regulation
Bridges highlighted that regulation itself is not the primary barrier. In England and Wales, regulators often treat AI oversight similarly to trainee supervision, focusing on critical review points rather than constant oversight.
The more significant bottleneck is professional indemnity insurance. With minimum cover requirements of £3 million per claim, insurers must be comfortable underwriting AI-driven services. Data, therefore, becomes central. Firms that can demonstrate statistically reliable outcomes may find insurers increasingly receptive.
Over time, actuarial evidence may allow certain workstreams to be largely automated with confidence.
The Economics of AI-First Models
A key advantage of starting from scratch is cost transparency. New AI-native firms can draw a direct line between automation and margin. Established firms, however, struggle to attribute savings when AI tools are used inconsistently across large teams.
Being first to market is not necessarily an advantage. With technology evolving rapidly, early movers risk building ahead of client readiness. Legal adoption remains shaped as much by trust and habit as by capability.
Culture, Leverage, and Cognitive Over-Load
AI is also reshaping the human experience of legal work.
If lower-value tasks are automated, senior lawyers may find themselves spending entire days on high-level strategic thinking. Whilst appealing in theory, this creates significant cognitive strain. The removal of routine tasks does not necessarily reduce pressure; it may intensify it.
At the junior end, increased automation may not eliminate bottlenecks. Traditional leverage models with many juniors funnelled into fewer supervisors still apply. Firms may need to rethink pyramid structures entirely if AI meaningfully alters workload distribution.
The Next Five Years
Both speakers expect significant growth in AI-native firms within the next few years. The true inflection point may come when established partners leave major firms to build AI-first ventures independently.
Rather than incremental tool adoption, AI nativity represents a structural shift in how legal services are conceived, delivered, and priced. The firms that succeed will be those that begin not with technology, but with a clearly defined, commercially meaningful problem.
