The Platform Vs Specialist Debate is Asking the Wrong Question

By Katja Nikolaus, Co-Founder, JUNE.

Two forecasts. Two opposing verdicts. Published within weeks of each other, each delivered with equal confidence.

The first: the market is pulling away from broad platforms toward domain-specific AI built for discrete legal subfields. Teams that specialize will outperform those trying to retrofit general systems (National Law Review, 85 Predictions for AI and the Law in 2026). The second: the opposite. Point solution adoption is plateauing as users hit solution fatigue – individual AI tools are approaching their ceiling. Stop collecting tools. Build infrastructure instead. (Artificial Lawyer Predictions 2026)

Many people deciding where their 2027 budget goes are holding both on the desk right now. The natural response – pick a side – is exactly the mistake.

I don’t think these forecasts contradict each other. Both describe something real. Holistic platforms struggle with the depth a single legal area demands. I have watched horizontal tools flatten the specifics of a regulatory regime until what remained was useless to the people actually running the cases. Point solutions create their own problem: a drawer full of sharp instruments with no view across any of them.

The error isn’t in either diagnosis. Both are valid. It’s in the conclusion that follows – the assumption that you have to choose. That assumption made sense when legal AI meant helping one lawyer work faster on one matter. It makes far less sense when you look at what is actually happening inside large legal departments today.

Look closely and the shift turns out to be structural. The work itself has changed shape.

One event no longer produces one case. It produces thousands.

I spend a lot of time with legal departments at airlines, insurers and large enterprises. The pattern looks fundamentally different from a few years ago. A weekend of cancellations triggers several thousand EU261 claims by Monday morning. A data incident becomes a wave of parallel requests with a multitude of deadlines. A regulatory inquiry turns into hundreds of near-identical files, each with its own status, its own escalation path, its own clock running. The legal question inside any single one of them is often straightforward. Run them together, at that volume, and the difficulty becomes one of visibility, whether anyone can see across all of them at once.

Before going further, a boundary worth drawing, because this argument doesn’t apply to every corner of legal practice. Teams handling individual contract disputes, working through a negotiation strategy with a client, or structuring a complex transaction are dealing with single matters where legal depth is everything and volume is irrelevant. For that part of the market, the platform-versus-specialist debate may actually land correctly.

But that is a structurally shrinking share of what large legal departments face day to day. What is growing, year on year rather than in cycles, is the other kind: mass proceedings, serial claims, regulatory waves that turn a single trigger into thousands of near-identical files. That work obeys a different logic entirely. And neither platform nor point solution has offered a convincing answer to it.

The data confirms the pressure is real and building.

The CLOC 2026 State of the Industry Report documents a structural productivity gap: workload demand is surging – particularly in regulatory compliance and cybersecurity – while budget and headcount growth have flattened. Only 32 percent of legal departments expect any increase in attorney headcount. More volume, same team. That cannot be solved by working harder. It can only be solved by operating differently.

This is precisely what the platform-versus-specialist debate misses.

A specialist tool that handles aviation law beautifully still leaves you operationally blind across four thousand parallel EU261 files. A broad platform that gives you a shared view but loses the substantive depth of each legal regime leaves you organised and wrong. Neither answers the question that has actually shifted inside these departments: not how to handle one matter well, but how to run four thousand of them without missing a single deadline.

The forecasters noticed this, almost in passing. In-house counsel will discover through AI that a large chunk of their workload is operational or business logic – not legal risk assessment. That deserves to be the headline, not just a footnote. At volume, legal work becomes an operations problem.

The adoption numbers reflect an industry reaching for something it hasn’t quite found yet.

The ACC/Everlaw survey found that GenAI use among U.S. corporate legal respondents more than doubled in a single year, from 23 to 52 percent, while Europe led globally at 61 percent adoption.

Adoption at that speed signals something bigger than a better way to draft. Organisations are realising their volume has already outgrown how they were set up to work.

Separate Wolters Kluwer/Legisway figures point in the same direction: 56 percent of European legal departments use general tools such as ChatGPT, while only 14 percent have implemented specialised legal AI solutions. Fast adoption, shallow depth. What many departments have built is, at best, digitised improvisation: more tools, no system, no visibility across the whole. That is not the progress the numbers suggest.

What’s missing sits between the two, the connective layer that neither a bigger platform nor a sharper point tool provides.

An operating system has always been a third thing: a common layer that carries specialised applications while holding control across the whole. Translated to high-volume legal work: you need deep, vertical handling of each specific regime – EU261, data protection, regulatory enforcement, subrogation, and the list grows with every new wave of mass proceedings – all of it carried on a shared process layer that holds intake, deadlines, documents and reporting across every one of them. Depth where the law lives. Control across the whole. That combination is what high-volume legal work has lacked – and it is the gap we built JUNE to close.

That sounds like a technology promise. It is actually an organisational one. The CLOC report makes the underlying failure clear: the challenge is one of governance: senior legal professionals tied up in high-volume routine work instead of the matters that need their judgment.. That is the real cost of the status quo. The missing layer between legal substance and operational reality.

The interesting story of 2026 is not which side of the platform-versus-specialist argument wins.

It is that the argument is being settled by the volume itself – quietly, without announcement, inside the departments that have understood that high-volume legal work still demands legal depth, plus an operational layer alongside it. Those departments are rebuilding now. The others are still buying tools.

Turns out, the forecasts were both right. They were just answering a question the work had already moved past.

–

To learn more about what JUNE can do for you see here.

—

[ This is a sponsored thought leadership article by June for Artificial Lawyer. ]

—

(Conference advert)

Come and join us in New York and London this November at Legal Innovators!

Legal Innovators New York – Nov 17 and 18.