Over 80% of enterprise AI projects fail, not because of the technology, but because organisations skip the foundation: literacy, workflow redesign, and leadership commitment. The difference between "having AI" and being AI-empowered is where competitive advantage is made or lost.
Lawyers are trained to deconstruct everything to its fundamental parts. To understand every clause, every implication, every angle. To use precise language. To leave nothing to chance.
Which is exactly why most of them are struggling with AI.
At the centre of every AI sits a black box. Non-deterministicProducing different outputs from the same input each time. Unpredictable. Impossible to fully understand. You can ask the same question twice and get different answers. You can craft a perfect prompt and still receive confidently wrong output.
This is the most uncomfortable technology the legal profession has ever faced.
And it is not slowing down. This technology evolves faster than anyone can follow, and it spreads instantly. When AnthropicAI safety company, maker of the Claude AI assistant recently released a legal-specific plugin for ClaudeAnthropic's AI assistant, used widely in professional services, every lawyer on the planet could access it within minutes. Low cost. Sometimes free. Never in human history has such a disruptive technology proliferated so rapidly. And unlike any technology before it, this one talks. It stirs something primal in us. It challenges what we thought made us unique.
Yet this discomfort is precisely why legal professionals are uniquely positioned to master it. The same skills that make lawyers cautious — critical reading, rigorous verification, attention to detail — are exactly what AI requires. The question is not whether lawyers can use AI. It is whether they will start before the gap becomes uncrossable.
The real problem: installed does not mean used
A few months ago, we sat with the leadership team of a major European corporate. When we asked about AI, they beamed with pride: "We've rolled out Microsoft CopilotMicrosoft's AI assistant integrated into Office 365 applications across the organisation. Everyone has access. We're covered."
Then we asked the follow-up questions. How many people actually use it? What has changed in how work gets done? What impact are you seeing on KPIsKey Performance Indicators — metrics used to measure success?
Silence. Uncomfortable glances.
This is not an outlier. It is the pattern.
20%
of organisations actually measure ROI on AI investments
Thomson Reuters, 2025
18%
have implemented a proper AI governance framework
LEGALFLY, 2025
70%
of legal teams have received no AI training whatsoever
Consilio, 2025
We have seen this before. Ten years ago, the legal industry talked endlessly about legal tech. Which platform to choose. Which features mattered. And then almost nobody actually used it. The same analysis paralysis is happening again, just with higher stakes and faster timelines.
The problem is not the technology. It is the approach. Organisations treat AI like any other software procurement: hand it to IT, evaluate vendors, roll it out, check the box. But AI is not another tool to be deployed. It is a foundational technology, like electricity, like the internet, like mobile communication. It is the infrastructure layer upon which entirely new ways of working will be built.
Why most AI initiatives fail
The statistics are sobering. Over 80% of enterprise AI projects fail, twice the failure rate of traditional IT projects.3RAND Corporation / Gartner AI Project Studies, 2024–2025GartnerLeading technology research and advisory firm predicts that at least 30% of generative AIAI that creates text, code, images, or other content projects will be abandoned after proof of concept.4Gartner: "30% of GenAI projects abandoned after PoC," July 2024 On average, only 48% of AI projects ever make it into production.5Gartner AI Production Survey, 2024
Three fundamental mistakes keep repeating.
Tool-first thinking
Companies buy AI tools before understanding the foundation. They pay premium prices for capabilities they could build themselves with a €24/month subscription, if only they understood how. Legal tech vendors are happy to sell you solutions. But without AI literacy, you cannot evaluate what you actually need versus what sounds impressive in a demo.
A mid-size corporate legal team spent €150,000 on a contract analysis platform. Six months later, usage data showed only three people had logged in more than twice. The problem was not the tool. It was that nobody knew how to integrate it into their actual workflows.
Organisations that see significant financial returns from AI are twice as likely to have redesigned their workflows before selecting tools.6McKinsey Global Survey on AI, 2024 Not after. Before. The technology serves the transformation, not the other way around.
Ignoring the human factor
Within The Legal Model, we bring experience from more than 70 enterprise AI transformations. One pattern stands out above all others: the higher you go in a company's hierarchy, the less direct experience people have with AI.
This creates a dangerous disconnect. Leaders who have never prompted, who have never experienced AI's capabilities and limitations firsthand, are making strategic decisions about AI adoption. They approve budgets, set timelines, and define success metrics without understanding what they are dealing with.
"I was terrified this would make me obsolete. Now I see it makes everything I know more valuable."
Senior lawyer, 30 years of experience, after a TLM workshop
The fears are real. Will I lose my job? Can I trust this? What about ethics? What about the billable hour? These are not obstacles to dismiss. They are signals to address. Every AI initiative that skips the emotional work fails. You cannot transform how people work if you do not first transform how they feel about the change.
Stopping at "faster"
Most organisations treat AI as a speed boost. Same processes, just quicker. Draft this contract faster. Review this NDANon-Disclosure Agreement — a contract protecting confidential information faster. Summarise this document faster.
That is not transformation. That is optimisation. And it misses the entire point. The real question is not "how do I do this faster?" It is "what can I do now that was impossible before?"
What transformation actually looks like
A senior IP lawyer attended one of our workshops. Afterward, he said something that stuck with us: "I just realised I'm also a large language model. I read texts. I remember knowledge. I bring cases into context with laws. And you just showed me that machines can do this part really well."
Then he paused. "So I can now think about how to create more value for my clients with the rest of my time."
That is the question. And the answer is not about doing the same things faster. It is about becoming something different.
Lawyer A — same day, same skills
Spends six hours drafting a complex supply agreement and two hours reviewing an NDA. End of day: two tasks completed.
Lawyer B — AI-empowered
Spends six hours building an AI assistant that drafts supply agreements in 30 minutes (with human review). Then builds an NDA triage agentAn AI system that sorts and prioritises incoming requests for sales. End of day: two systems that handle hundreds of tasks.
This is not about being 20% faster. It is a fundamental shift in what a legal professional does. From document producer to system architect. From bottleneck to enabler. From reactive to proactive.
We call this new role the Legal Solutions Architect: the person who bridges legal expertise and AI capability, who designs the systems that multiply their team's output. Not every lawyer needs to become one. But every legal department needs at least one, and every lawyer needs to understand what this role enables. The Legal Solutions Architect does not replace legal counsel. They transform what a legal team can deliver.
The productivity gains are measurable. Thomson Reuters' 2025 Future of Professionals report found that AI-empowered legal professionals save an average of 240 hours per year, equivalent to $19,000 in value per professional.8Thomson Reuters Future of Professionals Report, August 2025 GC AI's January 2026 study of in-house legal teams found they reclaim 14 hours per week through AI assistance.9GC AI Study: In-House Legal Teams, January 2026
But the real transformation is not about hours saved. It is about what you do with them. Being a General Counsel, imagine what is possible: your colleagues in sales can actually talk to contracts, asking questions and getting instant answers without waiting for your legal team. Your standard legal processes run themselves, flagging only what truly needs human judgment from your legal department. Your legal professionals spend their time on work that requires wisdom, creativity, and human connection. The parts of the job they actually went to law school for.
The gap is growing exponentially
AI is not slowing down. Every month, the technology becomes more capable. Every month, the gap between AI-empowered legal professionals and those still waiting widens. This is not linear growth. It is exponential.
11%
UK legal professionals using AI
July 2023
41%
14 months later
Sept 2024
55%
using AI daily or hourly
Jan 2026
95%
expect AI central to workflow
2030 forecast
Someone who started seriously working with AI two years ago is not just ahead. They are operating in a different category entirely.
We work with AI every day. We follow every development. And honestly, we have FOMO too. We cannot keep up with everything. Nobody can. If that is true for us, imagine what it is like for someone who has not started yet.
The window to build foundational AI literacy is open now. It will not stay open forever. Not because AI will "take over" (it will not). But because the professionals who master this technology will simply operate at a different level. They will win the clients, attract the talent, and set the standards.
Three principles for getting it right
We founded The Legal Model because we saw too many legal departments failing with AI. Not because they lacked budget or technology, but because they lacked the right approach.
01
Understanding first, tools second
We do not start with tools. We build AI literacy from first principles. How do these systems actually work? What can they do well? Where do they fail? Only when you understand the foundation can you make intelligent decisions about which tools, if any, you need.
This is not slower. It is faster. Because you stop wasting money on solutions to problems you do not have.
02
Human in the loop, always
You cannot trust AI. Full stop. You need to trust yourself.
AI is a tool: extraordinarily powerful, but still a tool. The approach is to use it the way you would use a brilliant colleague who sometimes confidently says wrong things. Verify everything. Build in checks. Maintain final judgment. AI-transformed content, never AI-generated content. There is a critical difference.
03
Capabilities, not dependencies
The goal is that you do not need us anymore. Build your team's capabilities, not your dependency on vendors or consultants. When we finish working with a legal department, they have the skills to continue growing on their own. They have workflows they built themselves. They understand what they are doing and why.
That is transformation.
Don't just adapt. Define the model.
The legal profession stands at a fork. One path leads to more tools and more subscriptions nobody uses. The other leads through temporary discomfort toward genuine transformation.
The technology will keep advancing, whether you are ready or not. You can use all the arguments you have, but the gap will keep growing. The legal professionals who master this will keep pulling ahead. The only question is: when do you start?
Ready to close the gap?
Find out where your legal department stands — and what it takes to get ahead.
A former lawyer and interim legal counsel who built House of Lawyers over eight years, connecting with over a thousand General Counsels and Heads of Legal across the Netherlands. She knows what legal leaders actually struggle with — not from surveys, but from thousands of conversations.
Chris Kwiatkowski
Has led over 70 enterprise AI transformations across insurers, pharma companies, and industrial holdings. More than 14,000 people have gone through his workshops. He brings the transformation methodology that actually works.
Yannick Bakker
A practising lawyer who uses AI every single day. He was talking about the future of legal work before ChatGPT existed. He combines deep legal substance with design thinking — understanding not just what is legally correct, but what actually helps people.