AI is no longer an experimental tool in the legal profession. It is becoming infrastructure, reshaping how lawyers work, price services and define their role, according to Tom G. Martin, founder and CEO of intelligent automation company LawDroid.
In an interview with The AI Innovator, Martin described a shift already underway: firms are moving past early experimentation with generative AI into practical, embedded use. “It’s not new. It’s not experimental,” he said. “It’s really become part of firms’ infrastructure.”
That transition is accelerating a broader change in how legal work is performed — and how much of it gets done.
But Martin also warned that AI’s efficiency gains are creating what he calls a “hyperproductivity trap,” where lawyers use automation to take on more work rather than reduce workload. The effect mirrors Parkinson’s law, where work expands to fill available time. Martin wrote about it a year ago, and the trend was later supported by research published in Harvard Business Review.
“The trap happens if you don’t set boundaries,” Martin said, arguing that the issue is cultural, not technological.
That cultural dimension runs through much of AI’s impact on the legal sector. While tools are improving rapidly, firms that fail to adapt risk falling behind. Martin compared the shift to the disruption of Blockbuster by Netflix, noting that firms still hesitating due to concerns about accuracy or reliability may miss the larger structural change underway.
The more fundamental transformation, he said, is how lawyers define their work. AI is forcing a distinction between tasks and purpose. Routine activities such as drafting contracts or reviewing documents are increasingly automated, but the core function of law — solving client problems — remains.
“Too many lawyers have identified with the task,” Martin said. “If they identify with their purpose, which is solving problems, then they’ll feel invigorated.”
This shift is particularly evident with the rise of agentic AI, systems that can perform multi-step tasks and make decisions. While such systems can execute workflows, Martin emphasized that human judgment will remain central.
“Decisions are different than judgment,” he said. “Judgment implies a human element.”
That distinction is shaping how law firms may evolve over the next few years. Rather than operating as collections of manual processes, firms are likely to integrate AI across research, document analysis and case management. AI will not be a standalone tool but embedded in everyday workflows, similar to existing software systems.
From billable hours to outcomes
Pricing models are also expected to shift. As AI reduces the time required for many legal tasks, the traditional billable hour faces increasing pressure. Martin said the industry is moving toward outcome-based pricing, where clients pay for results rather than time spent.
“People don’t care about how it happens,” he said. “The outcomes are what our clients care about.”
At the same time, AI may enhance the quality of legal services rather than simply reduce costs. By automating time-consuming document review, lawyers can focus more deeply on analysis and strategy, potentially delivering better results.
The technology is also reshaping legal research. With the growing availability of open case law databases and AI-powered tools, traditional platforms face new competition. While Martin does not expect paid services to disappear, he said the market is expanding with lower-cost and more flexible options.
“There’s more options now,” he said, pointing to new tools built on open legal data.
Decisions are different than judgment. Judgment implies a human element.
Concerns around accuracy, hallucinations and data privacy remain, particularly with retrieval-augmented generation systems that combine proprietary data with large language models. Martin said current systems are “good enough” but not perfect, noting ongoing research aimed at improving reliability.
Within a year, he expects many of these issues to be largely resolved as models improve and validation techniques mature.
On ethics, Martin argued that existing legal rules are sufficient if properly enforced. High-profile cases of AI-generated errors, including fabricated citations, reflect human oversight failures rather than gaps in regulation.
“You take responsibility for what you file with the court,” he said, referencing Rule 11 obligations.
The larger challenge, he suggested, is not compliance but adoption. Education alone will not drive change. Firms must also adjust incentives, such as offering billable credit for AI-related work, to encourage lawyers to engage with new tools.
Legal education is beginning to reflect that reality. Martin said AI should not be treated as a standalone subject but integrated across the curriculum, as it will underpin nearly all aspects of legal practice.
“This is the world that these law students are going to be working in,” he said.
For practitioners navigating the transition, Martin pointed to a simple but recurring theme: curiosity. In a field facing rapid technological change, he said, maintaining a willingness to explore and adapt may be the most valuable skill.
“Curiosity is a superpower,” he said.
Be First to Comment