When Legal Production Is No Longer Scarce: AI and the New Economics of Law

By Lindsay O'Connor, General Manager Content Pacific, LexisNexis^®

This article draws on themes explored in a panel discussion on ‘AI and the Transformation of Legal Economics’ at the LexisNexis AI Summit, where Lindsay O’Connor moderated a conversation with Robyna May, Chief Information Officer at McInnes Wilson Lawyers; Mark Love, Legal Director, Business & Corporate at BAL Lawyers; and Indrajeet Aich, Acting General Counsel at Challenger Limited.

At the LexisNexis AI Summit, I recently had the opportunity to moderate a discussion on a question that I think is becoming increasingly important for both law firms and in-house legal teams: what happens to the economics of legal services when the cost, speed and scale of producing legal work begin to change?

In this article

• From individual productivity to a different operating model
• Who captures the productivity dividend?
• Clients have access to AI too
• Could AI grow the legal market?
• The talent question is more complicated than headcount
• What becomes scarce?
• Trust has an economic value
• The choices we make now

For the past few years, much of the legal industry’s conversation about generative AI has focused on productivity. Can AI produce a first draft faster? Can it accelerate research, review contracts or summarise large volumes of material? Can lawyers use it to reduce the time they spend on repetitive work? Increasingly, we know that it can.

What interests me now is what follows from that.

As AI moves beyond individual tasks and becomes embedded across research, drafting, review, internal knowledge, legal content and matter workflows, its impact starts to extend well beyond the productivity of an individual lawyer. It begins to affect how legal work is designed, who performs it, how it is staffed, what it costs to produce and, inevitably, how clients expect to pay for it.

From individual productivity to a different operating model

Most organisations began their AI journey with relatively discrete use cases: research, summarisation, first drafts and document review.

Those applications are valuable, but they largely improve an existing process. The more significant shift comes when these capabilities start to connect.

Research feeds drafting. Drafting feeds review. Legal content is combined with organisational knowledge and data. Repeatable processes become structured workflows. Defined AI skills can carry out individual parts of a legal task, and increasingly, agentic capabilities can coordinate multiple steps within it.

At that point, the opportunity is no longer confined to doing the same work faster; organisations can start to redesign the work itself.

McKinsey's recent analysis of the Australian legal market estimated that, even under a conservative scenario, between 15 and 25 per cent of total net legal hours could be affected by AI-enabled automation or augmentation over the next five to seven years.

However, I don’t believe that should be read as a prediction of a corresponding reduction in lawyer numbers. It is better understood as an indication of the scale of work that may be performed differently.

For law firms, this sits alongside a business model that has historically been built around scarce professional time. Rates, utilisation, leverage and career structures all reflect, to varying degrees, an assumption that producing more legal work requires more human effort.

AI begins to loosen that relationship.

Australian firms are already experimenting with what a different model might look like. Some have built multidisciplinary innovation functions bringing lawyers together with technologists, knowledge specialists, project managers, developers and data professionals. Others are developing secure internal AI environments and embedding AI into client-facing services rather than treating it purely as a lawyer productivity tool.

The selected technology is arguably less important than the operating model being built around it. Broad access to an AI assistant may improve individual efficiency; sustained transformation requires workflow design, knowledge infrastructure, governance and clarity about how the resulting capacity will be used.

Who captures the productivity dividend?

This leads to one of the more difficult commercial questions.

If work that previously took ten hours can be delivered to the same or a better standard with five hours of human effort, where does the economic benefit go?

• A client could reasonably expect some combination of lower cost, faster delivery or greater predictability.
• A firm may have made significant investments in technology, content, integration, training, workflow redesign and governance and will expect a return on those investments.
• The lawyer may use the capacity to manage more work or spend more time on complex judgments, negotiations, and client engagement.
• An in-house legal department may decide that previously outsourced work can now be handled internally.
• Technology and content providers increasingly form part of the infrastructure through which the legal service is delivered.

One of the most interesting themes to emerge from our Summit discussion was how differently the same productivity gain can look depending on where you sit in the legal ecosystem. A law firm may see capacity or margin; a client may see an expectation of lower cost or faster delivery; an in-house team may see an opportunity to insource work.

Indrajeet Aich, Acting General Counsel at Challenger Limited, put the client perspective particularly clearly: “I’ve always been of the mindset that value trumps time.” He also noted that some of the most valuable work, and the work for which he would be willing to pay a premium, may consume relatively few hours.

There is no obvious reason why the productivity benefit should belong entirely to one of those groups. In practice, competition, client expectations and the type of work involved are likely to determine how it is shared.

This is also where the discussion about the billable hour becomes more nuanced.

I do not expect hourly billing to disappear. For complex, novel or unpredictable matters, time may continue to be a useful component of pricing. There will also be work where the principal value lies in access to scarce expertise and senior judgment rather than the efficiency of production.

The pressure is more likely to arise in categories of work where AI and structured workflows materially reduce the human effort required. In those areas, clients will increasingly question whether elapsed professional time remains the most appropriate measure of value.

We may see a wider range of models as a result: fixed fees, portfolio pricing, subscriptions, managed services and, potentially, arrangements that link pricing more closely to outcomes or value delivered.

Robyna May, Chief Information Officer at McInnes Wilson Lawyers, drew an important distinction between time as a pricing mechanism and time as a measurement tool: “I think we conflate two things when we talk about the billable hour. One is how we actually bill our clients… The other element of the billable hour… is measurement.” As clients increasingly ask firms to demonstrate the productivity impact of AI, time recording may remain important even as pricing moves further towards value.

The more useful debate is therefore not whether AI ‘kills’ the billable hour, but where time continues to provide a sensible basis for pricing and where it does not.

Clients have access to AI too

There is another factor that makes this technology transition different from many previous waves of legal innovation. Clients are gaining access to increasingly capable AI at the same time as their external advisers.

For in-house teams, the economics are different. Their objective is not to generate more billable hours. It is to manage legal demand effectively, control external spending, reduce risk and support the organisation.

As AI capability improves, corporate legal departments will have more choices about where work sits. Some research, first-pass drafting, contract analysis, compliance activity and other repeatable work may be handled internally. Business users may gain access to better self-service tools. Legal departments may use AI to extend their reach without growing headcount at the same rate as demand.

This does not necessarily diminish the role of external counsel. It may, however, sharpen the distinction between work a client can perform itself and work for which it is prepared to pay a premium.

Deep expertise, strategic judgment, advocacy, negotiation, sector knowledge and accountability are much harder to replicate than legal production alone.

There is also an opportunity for firms to approach this proactively. Helping a client redesign a workflow may reduce the amount of routine work traditionally sent externally, but it can also deepen the relationship and reposition the firm around the areas where its expertise creates the greatest value.

That requires a broader definition of success than simply retaining the largest possible volume of work.

Could AI grow the legal market?

One of the assumptions worth challenging is that there is a fixed amount of legal work and AI simply makes that work cheaper.

Legal demand has always been constrained to some extent by economics. Organisations do not review every contract, investigate every emerging risk or seek advice on every business decision because doing so would be prohibitively expensive. Individuals and smaller businesses leave many legal needs unmet for the same reason.

If the cost of legal analysis and production falls, some of that suppressed demand may become economically viable:

• An organisation could monitor regulatory change continuously against its own obligations rather than relying on periodic reviews.
• A legal team could analyse an entire contract portfolio rather than a sample.
• Potential risks could be surfaced earlier.
• Legal guidance could be delivered closer to the point at which a commercial decision is made.
• Smaller businesses could gain access to legal support that would previously have been disproportionate to the value of the issue.

This is what I think of as the Legal GDP question.

If AI lowers the cost of producing legal work, do we simply spend less on legal services, or does the total amount of legal capability consumed increase because more of it becomes affordable?

The most likely answer is a combination of both.

There are already early signs that increased productivity does not necessarily translate into reduced demand for lawyers. Robyna May told the Summit that McInnes Wilson was “not seeing a reduction in the number of people that we are employing”, adding that “we’re almost seeing an increase in legal work rather than a reduction of it.”

Some existing work will become cheaper. Some tasks will no longer justify the same level of human involvement. At the same time, new categories of preventative, continuous and scalable legal services could emerge.

For firms and legal departments, that creates a strategic choice. The productivity created by AI can be treated principally as a cost-saving opportunity, or it can be used to broaden the services offered, improve risk coverage and address legal needs that previously went unmet.

The talent question is more complicated than headcount

Any discussion about the economics of AI must also consider how lawyers develop.

A significant amount of junior legal work is repetitive. That makes it an obvious candidate for AI assistance or automation. Yet the fact that work is repetitive does not mean it has no developmental value.

That does not mean every traditional training task should be preserved. Robyna May offered a useful challenge to this during our discussion: “We need to be careful about being too romantic about the past and how we learned legal skills. But we do need to think very critically about how we ensure that critical judgment.”

Researching cases, reviewing documents, conducting due diligence and preparing early drafts have traditionally helped junior lawyers build legal knowledge, pattern recognition and judgment. Senior lawyers can often complete those tasks more efficiently, precisely because they have spent years doing them.

There is a positive version of the AI story. Junior lawyers may be able to learn faster, work across a wider range of matters and spend less time on mechanical production. AI may provide better starting points and create more opportunities for supervision and discussion of the difficult issues.

There is also a risk that needs to be taken seriously: a lawyer supervising an AI-generated output needs enough substantive expertise to recognise when something is subtly wrong. If too much foundational work is removed without replacing its developmental function, the profession could create a gap between apparent capability and genuine expertise.

The issue is therefore broader than whether firms recruit fewer graduates.

Law firms will need to think deliberately about how professional judgment is developed in an environment where repetition may play a smaller role in production. Apprenticeship, simulation, structured learning and supervision may become more important rather than less.

The long-term talent challenge may not be that AI prevents people from becoming lawyers. It may be ensuring that AI-enabled lawyers still acquire the depth of expertise required to become excellent senior advisers.

What becomes scarce?

Perhaps the most interesting economic effect of AI is the way it changes scarcity.

Historically, legal production has itself been scarce. Research takes time. Reviewing takes time. Drafting takes time. Expert professional labour is expensive and difficult to scale.

As AI reduces the marginal cost of parts of that production, economic value is likely to shift towards capabilities that remain harder to replicate:

• Judgment becomes more valuable: knowing which issue actually matters and when intervention is required.
• Trust becomes more valuable: being confident that an answer is grounded in reliable sources, can be explained and will stand up to scrutiny.
• Context becomes more valuable: understanding the client, its business, its sector, its objectives and its appetite for risk.
• Accountability remains valuable: someone must ultimately exercise professional responsibility and stand behind the advice.

And I would add strategic imagination: the ability to use new capacity to create a better service rather than simply accelerating the existing one.

This changes the nature of competitive advantage.

AI may also change some of the traditional advantages associated with firm scale. Mark Love, Legal Director, Business & Corporate at BAL Lawyers, observed that mid-sized firms are now “in a position where you can… compete with large research teams in a manner that you couldn’t before.” If sophisticated research and analytical capability become less dependent on the size of the team behind it, AI could alter competitive dynamics as well as individual matter economics.

If access to highly capable AI becomes commonplace, possession of technology alone will not distinguish one firm or legal department from another. The differentiators are more likely to sit around it.

The quality of the organisation's knowledge. The way it combines authoritative external sources with its own expertise. The sophistication of its workflows. The depth of its client and industry understanding. The effectiveness of its governance. The ability to use technology consistently across an organisation rather than through isolated pockets of enthusiasm.

Most importantly, there is the ability to decide where human expertise adds the greatest value and design the service accordingly.

Trust has an economic value

Governance is sometimes discussed as though it sits on one side of a tension, with innovation on the other. For legal AI, I think economics suggest something different.

A system that generates an answer in seconds but requires extensive checking, correction and supervision may produce far less value than its headline time saving implies.

That focus on quality rather than speed also came through strongly in the Summit discussion. Mark Love described one of the key advantages of AI as quality: “It gives you the opportunity to look at particular issues more thoroughly, more deeply and, most importantly, to test the material before it goes out the door.”

Productivity needs to be measured after the cost of verification, rework and risk is taken into account.

Reliable sources, provenance, explainability and appropriate human oversight therefore have a direct commercial role. They can reduce the time required to verify outputs, increase confidence in the workflows organisations are prepared to automate and enable AI to be used in higher-value contexts where an organisation would otherwise judge the risk to be unacceptable.

The same applies to governance. Well-designed controls should make it easier to scale AI responsibly across an organisation, rather than forcing each individual lawyer to make their own assessment every time they use it.

In a market built around professional confidence, trust is not separate from economic value. It is one of the things that allows that value to be realised.

The choices we make now

The traditional legal model is not going to disappear overnight.

The billable hours will remain. Lawyers will continue to sit at the centre of high-value legal work. Firms will still need to develop junior talent. Clients will continue to pay substantial premiums for outstanding advisers.

However, at the same time, the underlying economics are beginning to change.

For law firms and legal departments, there is a meaningful difference between using AI to improve the efficiency of the current model and using it as a catalyst to redesign legal services.

The first focuses on faster research, faster drafting and reduced production time. The second poses more difficult questions:

• What work should lawyers continue to perform themselves?
• What can be delivered through a workflow?
• What should move to the client?
• How should a service be priced when production costs fall?
• How should we measure value if hours become less important?
• How do we develop future experts?
• What legal needs can we now address that were previously uneconomic?
• And what will clients genuinely pay a premium for?

Those questions will not have the same answer across every organisation or every area of law.

There is also a broader opportunity here that goes beyond the economics of firms and legal departments. If AI can lower the cost of legal capability and make it available more widely, it has the potential to support greater access to justice as well as greater efficiency. That possibility connects directly with LexisNexis' longstanding commitment to advancing the Rule of Law.

As Robyna May said in closing, “I would hope we move into a world where access to justice is actually improved by the use of AI rather than being just another commercial barrier.”

That is an important test for the industry as this technology develops. The opportunity should not be measured only by how much faster or cheaper existing legal work becomes, but also by whether AI allows legal expertise, information and support to reach more people and address needs that the current model leaves unmet.

That is precisely why the next phase of legal AI is likely to be more interesting than the first.

Technology is becoming increasingly capable and increasingly accessible. The greater source of differentiation will be the choices organisations make about how to apply it, what they build around it and where they believe the future value of legal expertise will lie.

EXPLORE THE FULL LEXISNEXIS AI SUMMIT SESSIONS

Discover how LexisNexis Protégé™ can help your organisation connect trusted legal content, AI-powered skills and structured workflows to transform how legal work gets done.

Contact Us