One of the most persistent questions inside corporate legal teams sounds deceptively simple: how should we structure the function?

Historically, the answer was linear. As the business grew, so did the team. More contracts meant more lawyers. Increased regulatory pressure meant more senior hires. Complex matters were escalated to external counsel. The model was reactive and, for a time, effective.

That assumption no longer holds.

General Counsel today operate under sustained cost pressure, faster business cycles, and rapidly evolving technology, most notably AI. At the same time, expectations have shifted. They are no longer judged solely on risk management, but on their contribution to measurable business performance.

Against this backdrop, the question of structure begins to change. The question, therefore, is no longer how many lawyers are needed. It is how legal capability should be designed, so it can scale, operate efficiently, and align with the business.

The Design Problem Most Legal Transformations Miss

Much of the current conversation about AI in legal focuses on technology implementation: which tools to use, in what order, and how to drive adoption.

But there is a prior question that receives far less attention, that even well-implemented AI underdelivers if the underlying model remains unchanged. Work still flows the same way. Senior lawyers still absorb routine demand. The function continues to operate reactively because it was designed to.

The deeper challenge is not what AI makes possible within an existing structure. It is how AI should reshape the structure itself.

When you start there, the answers look quite different.

AI as a Structural Input, Not a Productivity Tool

AI is most commonly framed as a capability that makes legal work faster. Contracts are reviewed more quickly. Research is conducted more efficiently. Documents drafted with less effort. All of this is true, and the productivity gains are real.

But framing AI primarily as a productivity tool understates its significance. The more important effect of AI is not that it speeds up existing work, but that it changes where human expertise genuinely needs to sit.

When AI can handle document review reliably, the question is no longer how to do that work more efficiently. It is whether it should sit with lawyers at all.

This is a structural question, not a technology question. The shift has direct implications:

• How senior time is allocated
• How work is routed and governed
• How the function scales without proportional headcount

A legal function designed with AI as a structural input looks fundamentally different from one where AI is layered onto an existing model.

A Capability-Layered Approach to Legal Design

The most effective legal functions are moving away from thinking of themselves as a team and toward thinking of legal as a system of interconnected capabilities.

Legal work varies in complexity, risk, and volume. It requires different types of expertise, delivered in different ways, at different costs. A capability-based model makes this visible and allows it to be managed deliberately.

• At the top sits the strategy layer, led by the General Counsel, where legal aligns with the business, risk is interpreted commercially, and decisions are made about how legal capability should be deployed across the organisation.
• Beneath this sits the advisory layer: senior counsel providing high-value judgment on complex matters. Their time is finite and consequential. One of the defining characteristics of a well-designed legal function is how effectively this layer is protected from work that does not require it.
• The execution layer forms the operational core – contract negotiation, stakeholder support, day-to-day delivery. In many organisations, this layer absorbs a disproportionate share of the workload, often without the support structures needed to operate at scale.
• That support comes from the legal operations layer, which introduces the workflow governance, external spend management, and decision-relevant data that allow the function to run as a system rather than a collection of individuals.
• Closely linked is the technology and engineering layer – and it is here that I believe the most consequential transformation in legal design is currently taking place. More on this shortly.

Beyond this, flexible resourcing provides scalable capacity, while external counsel is used selectively for specialist expertise, rather than as a default extension of the team.

The Legal Engineer: The Most Underappreciated Role in the Modern Legal Function

If there is one role that is simultaneously the most important for legal functions right now and the least understood by those who lead them, it is the legal engineer.

Writing recently on the subject, Elly Meenan, Legal Operations Engineer at Wordsmith AI, offered a definition that cuts through the noise:

‘A legal engineer maps processes and then builds the solution themselves. Both things.’

That distinction matters. It is not enough to identify inefficiencies or define requirements. Legal engineers work hands-on with tools, designing and implementing systems that are actually used.

Their impact lies in ownership. The work is not done when a workflow functions, it is done when it is adopted.

The role of a legal engineer sits at the intersection of legal context, operational processes, technical architecture, and change management. And its absence is increasingly visible.

Critically, this role does not require a law degree. What it requires is enough legal literacy to understand how lawyers think, what they are afraid of, and why a process that seems straightforward will face resistance. That instinct can be developed. The ability to look at a broken workflow and see what it could become is harder to teach.

This is not a marginal improvement. It is a structural shift.

For any legal department of scale, engineering is no longer optional. It is foundational.

And the pipeline for this role is broader than what most GCs realise. As Elly Meenan of Wordsmith AI identifies, legal engineers enter the profession through three main paths: (1) from legal operations, where the remit has been quietly evolving toward engineering for years, (2) from non-legal backgrounds in product management, software development, and business operations, bringing building skills the legal world has historically lacked, (3) and from law itself, where the analytical instincts that make a good lawyer, interrogating a system, finding its edge cases, asking ‘but what happens when X?’, translate directly into workflow design when applied to tools rather than contracts.

The common thread across all three paths is not a credential. It is a threshold: can you map it, build it yourself, and own it over time? That is what separates legal engineering from legal tool use.

A General Counsel who understands this will hire deliberately for it. Those who do not will continue to wonder why their technology investments underperform.

Right-Sourcing as a Design Decision

Right-sourcing has long been understood as allocating work efficiently. In a modern legal function, it becomes something more fundamental: a design principle.

When AI handles repeatable work that does not require human judgment, and when a legal engineer has designed the systems that route and manage that work, right-sourcing decisions shift from questions of efficiency to questions of design.

• Where does expertise genuinely need to sit?
• What requires continuity and context?
• Where is defensible, experienced decision-making truly necessary?

These are a more useful set of questions. And a capability-layered model that is designed with AI and engineering in mind from the outset is far better positioned to answer them than one that allocates work based on the assumptions of a pre-AI, pre-engineering structure.

From First-Order to Second-Order Thinking

There is a deeper shift happening in how legal work is being understood.

For many years, the dominant focus was first-order improvement: reducing turnaround times, cutting costs, increasing throughput. These were valid priorities when volume was the primary constraint.

That is no longer the constraint.

Most legal teams are still optimising for speed, cost, and throughput, but increasingly, that is the wrong problem to solve. AI has compressed the front end of legal delivery. Drafting, review, and research are faster, more accessible, and increasingly commoditised. In many cases, the analytical and procedural dimensions of legal work are being absorbed by technology.

What remains is judgment, experience, and context.

The demand emerging across matters is not for general capability delivered quickly. It is for precision, highly specific expertise, continuity as situations evolve, and defensible decision-making once the routine work has been removed.

This is a different problem to solve. It is not about moving faster. It is about understanding where risk actually sits once speed is no longer scarce. A capability-layered model, designed with AI and supported by structural input and engineering capability, is positioned to address exactly that.

Implications for General Counsel

This evolution changes the role of the General Counsel. It means moving beyond the role of legal advisor to that of system designer to:

• Making deliberate decisions about how work flows through the function, where resources are deployed
• Hiring differently, not just for legal expertise, but for the operational and engineering skills that allow a modern legal function to build, adapt, and scale
• Collaborate with procurement, technology, and business

This is not a trivial change. It takes time, focused attention, and a willingness to challenge long-standing assumptions. But the benefits are significant. A legal function designed this way is more efficient, responsive, scalable, and better positioned to evolve alongside the business it serves.

Designing Legal for What Comes Next

Legal is not being replaced. But it is being redesigned. The organisations that navigate this most effectively will treat it as a design challenge, not a technology procurement exercise.

The forces driving this change (cost pressure, technological advancement, and shifting business expectations) are not going to reverse. If anything, they will accelerate.

The legal functions that will thrive are not necessarily those with the largest budgets or the longest list of platforms. They are those who have asked the harder questions:

• How should legal be structured, given what AI now makes possible?
• Where should expertise sit?
• Who in the function can build the systems that allow everyone else to operate at their best?

A capability-layered, AI-informed approach to legal design, one that takes engineering seriously as a core function, not an afterthought, offers clear answers. And in doing so, positions legal not as a function that reacts to demand, but as one that is deliberately designed to meet it.