The Legal Industry Is Chasing a 10–20% Gain. Here's How to Get 200%

I spoke at ACC Legal Ops Con last week in Chicago, and I opened with something that I think a lot of people in the room already knew but hadn't quite put into words yet.

Most legal AI today is focused on making lawyers faster. Better research tools, faster redlining, smarter drafting. And of course that makes sense. The system is under stress, lawyers are stretched thin, and the business wants answers asap… so you give the lawyers better tools and hope the throughput improves.

However, the reality is, you can do ALL of that and only move the needle 10 to 20%. 

Why is this the case? You’re optimizing the wrong part of the system.

The real gain — we're talking 50 to 200% — doesn't come from making lawyers faster. It actually comes from changing how work reaches them in the first place. 

The funny thing is, two industries already proved it before legal ever got there and the lesson is sitting right in front of us.

The Lesson From Outside Legal: Two Industries That Figured This Out First

Let’s have a look at hospitals. Emergency rooms were under enormous pressure to move patients through faster, so the instinct was to give doctors better endpoint technologies — better equipment, better tools at the bedside. And while those things are still valuable, when you look at what actually moved the needle on ER throughput, it wasn't any of that. It was medical triage. A structured system that helped manage what reached the doctors in the first place.

Customer support teams ran the same experiment. When AI first hit the industry, companies went straight to making their support agents faster and more efficient. And again, that was useful, but not transformative. The real lift came when they started deciding what even reached agents at all. Rules-based deflection/re-routing. Self-service. Handling it before it became someone's problem to solve.

The pattern is identical in both cases. A 10 to 20% gain from putting better tools in people's hands. A 50 to 200% gain from changing the system those people operate within.

When I look at legal — an industry dealing with the exact same pressure, the exact same backlog, the exact same "we need to do more with less" mandate — I have to ask: why are we still reaching for the same playbook that only gets us 10 to 20%?

The answer, I think, is that we haven't had a clear enough picture of what the alternative actually looks like in practice, until now.

What the Legal Operating Model Looks Like Today

More often than not, work comes into legal from every direction. Email, Slack, Teams, a tap on the shoulder, a call to someone's favourite lawyer… The business doesn't always know how to instruct lawyers, there's a lot of back and forth, and sometimes legal receives requests that shouldn't have gone to legal at all. I asked the room in Chicago if any of this sounded familiar and pretty much every hand went up.

And when I asked what the real pain points were, the answers came fast:

• Delays → A request lands in the inbox, sits there for two hours before anyone picks it up, then needs clarification, then the business takes two days to respond, then it goes back in the queue. Three days in and the work hasn't even started.
• No visibility → All of the activity is siloed in people's DMs, inboxes, and Slack messages. There's no full picture of what the department is actually doing, which makes resource allocation and proving legal's value nearly impossible.
• No load balancing → One lawyer is completely overwhelmed while another has capacity — and there's no way of knowing that, let alone acting on it.
• Duplication → Work gets picked up twice, two lawyers work the same matter and come back with different advice, and the business loses confidence.
• No prioritization → A five-minute task and a five-week task land at the same time with no way to systematically tell them apart.

Here's what I think is the most important insight buried in all of this: data is not a CLM problem or a system-of-record problem. It's an intake problem. It starts at the very point where work originates. And until you fix that entry point, everything downstream stays messy.

This is the operating model that most in-house legal teams are still running on. Tools get layered on top of it. But the model itself rarely changes.

Three Teams. Three Different Problems. One Answer.

I was lucky enough to share the stage in Chicago with three legal ops leaders who have each gone on their own version of this journey. And what struck me most was that they didn't all start from the same place. They came to the same answer from completely different pain points.

Ann Schlaffman, Managing Executive of Legal at Kyndryl, came to this same conference a year ago as an attendee. She sat in the audience and had a realization. Her problem wasn't that work wasn't getting done. Everyone found their way to their favorite lawyer eventually. The work got done. But Ann had nothing to show for it. No data, no visibility, no way to answer basic questions about what her team was actually working on or whether they were focused on the right things.

‍

‍

One year later, after implementing a Legal Front Door, she has the data she needs, the visibility to make staffing decisions, and a team that no longer operates out of a shared inbox. The operating model changed, and the data followed.

Midori Vasquez, Legal Operations Manager at Elastic, started from a completely different place. Instead of visibility, she had a contracting problem. The team went looking for a CLM, but since they couldn't find one that fit, they decided to build a modular alternative instead. Through this, she was able to redesign how contracting work entered and flowed through legal in the first place and automate the NDA process. The Legal Front Door was the entry point that made everything downstream possible.

‍

‍

Jeannine Moran, Global Head of Legal Operations and AI at Hitachi, came at it from a third angle entirely. Hitachi had moved to a shared services model, providing legal services to subsidiary companies and other Hitachi group entities across the globe. They needed one centralized entry point for every request that came into legal, regardless of where it originated or what it was. A single source of truth. For Jeannine, the Legal Front Door became the orchestration layer for everything legal does.

‍

‍

The results speak for themselves. Over 2,700 AI agent conversations resolved without ever reaching a lawyer. A global conflicts of interest disclosure process that brought in 1,000 submissions from around the world — 90% auto-approved below the risk threshold, zero attorney time required.

Three companies. Three completely different starting points. But the same answer every time: change the system that work flows through, and the gains take care of themselves.

What the Operating Model Change Actually Looks Like

So what does this actually look like in practice? Because I think this is where a lot of people get stuck. They understand the problem but the solution feels abstract.

Here's how I think about it.

The old version of the Legal Front Door was a portal. A place the business had to go to, learn how to use, and file a ticket. And that never really worked because nobody wanted to use it, and it felt like legal was putting up a wall between itself and the business. 

The modern version is different. It's not a destination — it's an AI-powered governance layer that lives wherever your business already operates. Email, Slack, Salesforce, Teams… it doesn't matter. The tool should live where your business lives.

When a request comes in, the system asks three questions in sequence:

1. Can this be self-served? If someone is asking a policy question, a guidelines question, or something that has a known answer sitting in your repository, the AI reads into that and answers it right where the question was asked. No ticket. No lawyer. Just an answer.
2. Is this a process? Because sometimes legal isn't just an answer, things happen before and after a request comes in. Can we kick off an automated workflow? Generate an NDA, process a conflict of interest disclosure, check a share trading window, route an outside counsel approval through the right checks and balances before a law firm ever gets engaged?
3. Does this actually need a lawyer? If it does, the AI captures the full context from everything that's already happened, auto-creates the matter, populates the fields, assigns it, and tracks it through to completion. No manual data entry. No status updates that never get made. Your data stays clean.

As Jeannine put it on stage, and I think this is the most important reframe in this entire conversation:

‍

‍

That's the shift. It's a structural change to how work flows through to legal — the plumbing, the architecture, the operating model itself.

Getting the Technology Right Is 20% of the Job

Honestly speaking, the technology is the easy part. 

Jeannine said something on stage that I think every single person in that room needed to hear:

‍

‍

Attorneys are risk-averse by nature. They're not naturally systems people. And yes — some of them will wonder what AI means for their role. But the mistake is leading with visibility and dashboards. Lawyers don't want to feel like they're being tracked. What they do want is to stop being the routing system and to stop spending their day on the five-minute tasks, status chasing, and back and forth that has nothing to do with legal judgement.

If you want to get your attorneys on board with adoption, frame it in a way that shows you’re protecting their time so they can be more strategic, better business partners, sitting at the table shaping deals before they land on someone's desk.

The same principle applies to every stakeholder group. For the sales team at Elastic, it wasn't a legal technology project. It was a deal accelerator. For the business at Hitachi, it wasn't an intake system. It was Hikari — their guiding light. The name, the story, and the framing all matters more than most people expect.

The Question Worth Asking

The results I heard on that stage in Chicago came from legal teams at very different stages of their journey. Some were early in building out their legal ops function. Others were further along. But what struck me was that the size of the team or the scale of the budget wasn't what determined the outcome.

What determined the outcome was asking a different question.

Not "how do we make our lawyers faster?" But "how do we change what reaches our lawyers in the first place?"

That question is worth 200%.

If you're a GC, a CLO, or a legal ops leader reading this and nodding along — chances are you already know the problem. You're living it. The good news is that the path forward is clearer than it's ever been, and there are people who have already walked it.

The only thing left is to take the first step.