The rise of the AI native lawyer

IN Partnership with

A former lawyer and data engineer explains why AI has become essential to legal practice and how LexisNexis Protégé keeps lawyers firmly in the driver’s seat

A MACLAREN racing car handed to someone who has only ever driven an automatic will not win many races. The power is there, but without the right setup and support, most of it goes unused. For Elizabeth Bing, a former lawyer and data engineer who is now a client development partner at LexisNexis New Zealand, that image sums up why buying a powerful AI tool and getting value from it are two different things. Her legal career taught her what a lawyer needs before trusting an output. As a data engineer, she implemented systems that had to comply with privacy and data laws in New Zealand, Europe and Japan. Every point at which data passed through had to be compliant, and that taught her how to build for trust at scale. Such a rare mix of skills shapes how she views the tools available to the profession today.

LexisNexis is a leading provider of legal, regulatory and business information, renowned for integrating advanced AI into its legal tools. Its AI-powered solutions – such as Protégé – enhance legal research, drafting and analytics by delivering precise, localised and context-aware results. Using natural language processing and machine learning, LexisNexis helps legal professionals streamline workflows, analyse case arguments and uncover insights from vast legal data sets. These innovations improve accuracy, efficiency and strategic decision-making in legal practice. Committed to responsible AI, LexisNexis ensures transparency and data privacy while continuously evolving its tools to meet the dynamic needs of legal professionals worldwide.

Find out more

“At LexisNexis New Zealand, I partner with law firms to maximise the value of Protégé by understanding their business needs and ways of working and shaping how Protégé fits into their practice”

Elizabeth Bing, LexisNexis New Zealand

“Lawyers give you judgement, engineers give you scale, and legal AI needs both,” says Bing. Few lawyers now have the option of ignoring AI. The open question is whether they can drive it well, given an understanding of how legal work is structured, how the tools work under the hood and how much control they keep.

From contracts to code Bing practised as a transactional lawyer, and the habits she developed transferred well into engineering. “Each profession complements the other. Law gave me attention to detail, a methodical approach and the ability to build trusted relationships with stakeholders,” says Bing. Engineering then gave her career the discipline. That real value starts with understanding the business need, and that a system only succeeds if the people it’s built for actually use and benefit from it. “I learnt how to build engineering systems at scale and manage change so that the systems we built were adopted. You build based on value and on understanding what stakeholders are trying to achieve, which means knowing the difference between wants and needs, then translating the business rules into coding logic that can be used at scale,” she explains. Both lessons carry into her current work, where a legal AI tool only succeeds if it delivers the value lawyers need to make it part of their daily practice. “At LexisNexis New Zealand, I partner with law firms to maximise the value of Protégé by understanding their business needs and ways of working and shaping how Protégé fits into their practice,” says Bing.

Seeing legal work as a pipeline Data engineers spend their days extracting information from where it sits, then cleaning the data so it is in a state where it can be used. “A lot of the data engineer’s work is literally being the plumbers of the world, trying to plumb data from A to B so that it can be used to create real insights,” says Bing. Law firms have no shortage of material to plumb. “We have things like file notes, calls, emails and client meetings. We’ve got a lot of unstructured data. We also have the formal contracts, or the case law we’re looking at for precedent,” she explains. Where a lawyer reaches for the IRAC structure to frame a problem, an engineer breaks the same job into stages and asks what happens to the information at each one. “There’s still the data at the end of the day, and how does that get extracted? Once it is extracted, in what form is it used? And how is that applied to legal judgement? Think of that transformation step as the legal judgement,” says Bing. The structure holds whether the practice has one lawyer or several. “A sole practitioner spends time doing everything themselves: extracting key facts, drafting from a template where possible and following up where information is missing. In large law, a team reviews hundreds of contracts in a due diligence exercise and populates a spreadsheet with the relevant findings. The structure is more elaborate, but it’s the same principle of information intake, extraction and application,” says Bing. Repeatability is the first test an engineer applies, and it carries a consequence lawyers may not expect. “Data quality might not seem important to a lawyer, but it actually is. When we are applying these processes, how can we ensure that each client matter is equally assessed and completed to the same quality, especially when we start using AI tooling? That quality is the value customers are paying us for. Otherwise, what’s stopping a customer from doing it themselves with their own AI tool?”

The AI arms race Pressure to adapt is arriving from the bench as well as the market. “Overseas, we are seeing that some courts in class-action lawsuits are directing the legal profession to use AI to assist in their preparation for mediation and trial. What this means is that in the future, it could have a real impact on recovering costs. For example, a lawyer may be unable to recover the full amount of their costs if it is determined by the court to be unethical because they were the only one reviewing everything manually when AI could have completed it faster,” says Bing. Budgets bring their own pressure. “Arguably, it’s become an AI arms race, where firms with larger budgets can afford to invest heavily in AI. That creates the argument that AI is essential because if you don’t adopt and adapt, you’ll be left behind,” she adds.

“Arguably, it’s become an AI arms race, where firms with larger budgets can afford to invest heavily in AI. That creates the argument that AI is essential because if you don’t adopt and adapt, you’ll be left behind”

Elizabeth Bing, LexisNexis New Zealand

At the same time, AI has democratised the playing field, with smaller firms gaining access to knowledge tools that can compete with the research teams of larger firms. “You have to separate access to the technology from the ability to use it well. Having the same model is not the same as having the ability to use it well, and that is where implementation of the tool, combined with the judgement of the lawyer, brings the real value,” says Bing. Search engines once made a similar promise, and the profession outlasted them. “When Google first came out, we had more information than ever at our fingertips, but that didn’t get rid of the lawyer. We still needed lawyers to be able to use their judgement, like in past times when the norm was going into libraries to find the hard copy, we still need to know what to look for, what is reliable and valid, and what is still current case law,” adds Bing. “Which case is the most relevant to the set of facts in front of me right now? That context still lies with the human, and AI does not yet have the capability to process it.” Putting the lawyer in the driver’s seat An agent, in engineering terms, is software that can perceive its environment and act on it with whatever tools it has been granted. “What distinguishes agentic AI is the ability to go forth and execute actions. The majority use AI purely through the chat function, and that does not make it agentic. For lawyers, this opens doors, whether it’s gathering data, automated data entry or managing a customer account. The potential economic value of agents is enormous,” she explains. LexisNexis New Zealand has built that capability into Protégé through AI-guided research and skills, pre-built or custom agentic tools that handle complex, multi-step legal tasks from start to finish. Protégé sets out a plan that the lawyer can confirm, change or cancel, shows its progress and the files it is using as each stage completes, and lets the user refine the result in plain language until it is ready for review. Two lawyers on opposite sides of the same dispute will start that plan from very different places.

Published 27 Oct 2026

“What I’d keep my own is the driving of the AI: deliberate intent, the critical thinking and knowing what I am trying to find out, then altering the prompt accordingly,” she says. For a profession built on quality and matter control, the design choice is deliberate. “It’s a powerful functionality, giving control back and letting the human be the driver in using AI tooling more efficiently.” More context, fewer hallucinations The large language models behind these tools are less mysterious than their reputation suggests. “There’s a common misconception that AI is a silver bullet, like a magic black box that we can put something into and expect it to spit out the desired outcome, but that isn’t the case. At its core, a large language model is an advanced statistical prediction model. Newer training methods have made these models far more capable. However, that does not change the fact that the quality of what comes out depends heavily on what goes in, and on a lawyer checking the result,” says Bing. Protégé is grounded in LexisNexis’ own content, including case law, commentary and practical guidance, a larger pool than a general-purpose model can draw upon. “What that does is lower the amount of hallucination,” says Bing. Protégé’s Vault and sub-vaults extend the same principle to a firm’s own files, bringing a legal professional’s documents and knowledge together with LexisNexis content in one secure workspace, with up to 50 vaults per user. “Context means more reliability. The more data available, the less hallucination. Think of a lawyer with 100 cases at hand versus 1,000 at their fingertips,” adds Bing. “From a data perspective, vaults also empower organisations to ensure the separation of matters and help give the AI the relevant context.” For example, litigators preparing a chronology for a complex matter can lose hours to data entry. “Rather than create one from scratch, a lawyer can use a skill in Protégé to create a chronology from numerous documents in a fraction of the time,” says Bing.

What lawyers should be doing now Bing offers a minimum target that every practitioner should aim for. “All legal professionals now require a baseline understanding of our privacy and data obligations to our clients, and of which AI tools and features are appropriate to use, bearing those obligations in mind. Is the data stored overseas? Is it being used to train models? Where is it processed? Any firm and any person engaging with AI needs to be aware of where their client data is going and how it is being used,” says Bing. Firms can take much of the guesswork out of those calls for individual fee earners. “Firms can provide an AI assessment framework, a risk matrix and policies that educate fee earners and empower them to decide which tools they can use while staying compliant with their privacy and data obligations,” she explains. “For lawyers who have yet to open an AI tool, start exploring the tools. Take time to reflect and identify your most time-consuming and repetitive tasks. Nine times out of 10, those can be automated,” says Bing. And for those who are worried that AI will simply homogenise their arguments into bland algorithmic fare? “From the get-go, it will not give the same answer, and it will start looking for cases in support of the proposition you’ve given it,” says Bing. How well a lawyer frames that proposition now shapes what the tool can find. “That’s why it’s so important for lawyers to understand how to give that context, to equip the AI to give the best answer based on what it has available to it because it’s an extremely powerful tool, and an AI’s output is determined by the quality of the input.” Built for legal professionals, Protégé helps firms work smarter and deliver consistent, high-quality outcomes for their clients. Book a free demo to see it in action.

AI adoption gains momentum across APAC in 2026

Source: LexisNexis 2026 APAC AI Sentiment Survey report

70% of New Zealand lawyers use specialist legal AI

88% of corporate legal teams say AI saves time on lower-value work

53% of APAC organisations now measure AI’s value

Source: LexisNexis Protégé

PLANProtégé sets out a plan the lawyer can confirm, change or cancel.

Lawyers stay in the loop

TRACKIt shows progress, and the content and documents it’s using, as each stage completes.

REFINEUsers refine the result in plain language until it’s ready for review.

“Say I’m acting for the employer who wants to restructure their business. The kinds of questions I’m going to be asking and prompting AI with are going to be very different compared to if I’m acting for the employee raising a personal grievance for an unfair process,” says Bing. The plan gives the lawyer full control of the path the AI goes down and, with that, the first option to intervene and correct. “If I’ve asked about a restructuring, it might say, ‘These are the four sub-questions I’m going to look for,’ the first being the leading case law under the Employment Relations Act 2000 on restructuring. Already the user knows it’s on the right track. If it said it was going to start looking at the Human Rights Act 1993 instead, the lawyer knows from the get-go, with their own knowledge and reasoning, that they don’t want AI going down that rabbit hole, and they can edit the prompt to put it on the right track,” explains Bing. Skills in Protégé can run a multi-step task from start to finish, but the direction the AI takes still must come from the person at the wheel.