AI can draft, analyse and automate, but it can’t replace human judgement. Natalie Kim explains why the next generations of successful lawyers will be those who embrace technology while sharpening the skills that machines can’t replicate, ahead of her upcoming Law Society webinar and masterclass.
After almost two decades working at the intersection of law and technology in the United States, New Zealander Natalie Kim has returned home with a clear view of where artificial intelligence is taking the legal profession. The opportunity, she says, is much bigger than doing today’s legal work faster.
When Natalie talks about artificial intelligence and the law, she’s less interested in whether AI can draft a contract than in what happens to the lawyer when it can.
Much of the conversation has focused on substitution: which tasks can AI perform and could it eventually replace work undertaken by lawyers?
Natalie believes that risks missing the bigger change. “I think we’re overestimating on output substitution,” she says. “Is it going to write the contract instead of me? Is it going to replace the associate?”
What lawyers may be underestimating is how work itself changes. As AI allows organisations to create information, analysis and requests at far greater speed, the genuinely scarce resources may become something much more human: a lawyer’s attention and judgement.
An Aucklander who spent almost two decades in the United States, Natalie worked in private practice and with technology businesses including Apple and Cisco. Most recently, she was Head of Legal at Seattle-based clean technology company Omnidian, where she also led AI transformation.
Now back in New Zealand, she sees an opportunity to think deliberately about what good AI-enabled legal practice should look like.
Natalie expects AI to become an “exoskeleton” around the lawyer, taking on appropriate parts of legal work. That doesn’t mean handing everything to a machine. Instead, human judgement moves upstream.
The lawyer increasingly decides how work should be done: what can safely be automated, what requires human review, where risk sits and where expertise is most valuable.
“AI is not the right solution for everything,” Natalie says. “The critical skill is knowing the difference.”
That requires lawyers to make visible the small judgements embedded in their expertise: knowing when a standard matter isn’t standard, what context matters and when an approach needs to change.
If AI takes over repetitive work traditionally given to junior lawyers, how will they develop that judgement?
Natalie sees an opportunity to rethink training.
As a junior mergers and acquisitions lawyer, she spent countless hours reviewing contracts and completing signature blocks. There was learning in the repetition, but it was largely incidental.
“We never really thought about that from a pedagogical or training perspective,” she says.
AI forces firms to be more deliberate. There remains enormous value in an experienced practitioner sitting alongside a junior lawyer and explaining not simply what something says, but how they think about it: what matters, where the risks are and when the usual rule doesn’t apply.
Human-to-human learning may become more important, not less.
Natalie has seen what this can look like in practice.
As sole legal counsel, she developed an AI process for reviewing non-disclosure agreements. It evolved from a reusable prompt into a tool that could check agreements, suggest edits and prepare responses.
The bigger change came when she redesigned the process so the sales team could use it directly, with guardrails escalating unusual or risky agreements to her.
The result, she says, was a 99% reduction in NDA review turnaround time.
The significance isn’t really the NDA. It’s the method: understand the decisions within the work, determine what AI can safely handle, build in human oversight and redesign the workflow.
“AI transformation cannot stop at doing the existing work faster; it needs to reimagine the work,” Natalie says.
That human oversight remains critical. AI can produce remarkably useful work and still be wrong. For Natalie, responsible use means designing review around the work and its risk rather than simply trusting the technology.
“No technology in the world is going to force humans to have integrity and thoroughness,” she says.
For New Zealand firms wondering whether they are already behind, Natalie’s message is reassuring. This is still early.
There is no settled model for the AI-enabled law firm. The more important divide, she believes, is between lawyers who understand how to direct and interrogate these tools and those who accept what the technology gives them at face value.
Clients, too, are increasingly arriving with AI-generated answers and asking lawyers: “Is this right?”
Rather than diminishing the lawyer’s role, Natalie sees an opportunity. The conversation can move to what context is missing, why an answer changes when that context is added and where professional judgement makes the difference.
AI can remove work that consumes time without making the best use of legal expertise. It can make services faster and more accessible and give smaller teams capabilities previously available only with greater resources.
But perhaps its greatest opportunity is to allow lawyers to spend more time on the things for which clients genuinely need them: understanding context, solving difficult problems, exercising judgement and building trusted relationships.
Natalie’s philosophy is that technology should ultimately direct the best human thinking to where it matters most.
The future is unlikely to be lawyers on one side and machines on the other. It is more likely to belong to practitioners who understand both: who know what technology can do, recognise what it cannot, and have the confidence and judgement to decide where each belongs.
For a profession built on judgement, that is not a diminishment of the lawyer’s role.
It may be an opportunity to put judgement and human interaction back at its centre.
AI is beginning to reshape how legal services are delivered, priced, managed and experienced by clients. The Law Society is responding to both the practical and leadership challenges with learning opportunities for lawyers and leaders.
Here’s a snapshot of upcoming learning offered by the Law Society including CLE in 2026.
Thursday 5 November 2026, 1-2pm. With Charlotte Price, Hannah Hardy-Hones and Colin Abernethy.
The most compelling case for technology in family law is not efficiency for its own sake, but the space it creates for deeper client relationships and human connection. In this practical session, participants will be introduced to pain point mapping, encouraged to examine their working day and explore where and which tools can help.
Wednesday 18 November 2026, 10-11.30am. With Natalie Kim and panellists.
This session connects the commercial and workforce shifts with the leadership decisions that firms and legal teams need to make now. Hear from Natalie and panellists as they examine the people, leadership and operating-model implications for firms, in-house teams and legal careers.
Thursday 12 November, 1-2pm. With Ali Dibbenhall and Elizabeth Bing from LexisNexis.
Leading legal and technology experts Ali and Elizabeth will examine the impact of AI on productivity, performance, wellbeing, and the evolving nature of legal work. Attendees will gain practical insights into where AI is delivering value today, where human judgement remains essential, and what sustainable AI adoption could look like for the future of legal practice.
Thursday 10 December 6-8pm. Auckland. With Natalie Kim.
For lawyers, the challenge is not simply understanding agentic AI. It is knowing how to move from experimentation to supervised, responsible use in real legal workflows.
This practical masterclass uses a crawl–walk–run framework to help participants understand their current level of AI capability, identify the next step, and apply that learning through a hands-on workflow design exercise.
While AI can support aspects of legal practice, lawyers remain responsible for exercising professional judgment, protecting client confidentiality, and verifying the accuracy and appropriateness of all content generated by AI.