Banner artwork by Shutterstock AI Generator / Shutterstock.com
This is the first in a series of conversations with senior legal professionals navigating AI adoption inside a large global enterprise, exploring what the transition looks like when you're responsible for hundreds of lawyers across dozens of countries.
Each article brings a different voice. Together, they offer what the vendor brochures don't: an honest account of what it takes to get this right.
When more than 7,000 legal professionals gathered in New York for Legalweek 2026 in March, the annual conference where law firms and in-house legal teams go to understand, buy, and operationalize legal technology, the buzz was impossible to ignore.
Artificial intelligence, once a cautious topic in legal circles, had moved to center stage. Mathias Gaertner, General Counsel, Company Secretary and Member of the Executive Committee of ABB, attended with a clear agenda: cut through the noise and figure out what AI adoption actually requires for a legal function of our scale, operating in over 100 countries.
What he brought back was a framework for thinking, which is far more useful than a shortlist of tools to buy. His pragmatic reflections provide insights for legal functions of any size and in any location across the globe.
The view from Legalweek: A market transformed
Spiwe L. Jefferson: You came back from Legalweek, the annual conference where law firms and in-house legal teams go to understand, buy, and operationalize legal technology, having attended sessions and walked the vendor floor. What was your overall read on where legal AI stands right now?
Mathias Gaertner: The market has clearly arrived.
Every larger company and law firm is testing, trying, and implementing AI solutions to some degree. What struck me most was how broad and fragmented the landscape still is, there are an enormous number of vendors, but very few offer anything close to an end-to-end solution. Most specialize.
We have moved past the question of whether legal AI exists and works. The current focus is which tools, for which tasks, and on what foundation. It's also worth noting that many legal AI tools use Large Language Models (LLMs) from providers like Anthropic or OpenAI, building a “legal layer” on top. The quality of the output, particularly for non-English or non-US law contracts, is a critical consideration, as it depends heavily on the data the model was trained on.
Q: With so many vendors competing for attention, how do you think about tool selection without getting overwhelmed?
Gaertner: I think about it the way I think about a good handyman's toolkit.
You need more than one tool to build a house. A hammer doesn't replace a drill. And all of them still require someone skilled enough to pick up the right one and judge the result.
For legal teams, that means resisting the temptation to find a single platform that does everything.
The better question is: what are the distinct tasks we perform most often, which tools do each of those tasks best, and can we integrate them into a coherent workflow? At ABB, for example, we have established a cross‑functional expert group covering Digital and Intellectual Property, active across all business units. It supports both the development and protection of AI solutions — currently some of the most exciting roles at our company. At the same time, we are investing in AI as a tool to help our teams work better and faster. It will significantly enhance the agility and creativity of our Legal team.
Q: What about vendor longevity, how do you factor in the risk that a tool you invest in today might not be around in three years?
Gaertner: It's one of the most important questions that doesn't get asked enough.
Many of the tools at the conference are early-stage startups funded by venture capital. We cannot spend a year configuring a platform, uploading templates, building playbooks, training teams, and then watch the vendor disappear or get acquired.
We want the tools we invest in to have longevity. Vendor sustainability is part of the due diligence, full stop.
Data readiness: The foundation that comes before everything
Q: Beyond the tools themselves, what was the most important insight you took away from Legalweek?
Gaertner: Without question: data readiness.
Before you can get meaningful value from any AI tool, your underlying data must be in order. Contracts and other legal documents need to be available in a common place, in a common format. If your data are scattered, siloed, or inconsistently structured, your AI output will reflect that chaos. Garbage in, garbage out applies more in legal AI than almost anywhere else. In parallel, we have reduced our Group Governance complexity at ABB by 75 percent: we streamlined our Authorization Tables, Policies, and Procedures to make decision-making faster and accountability clearer, which also supports data readiness.
Before you can get meaningful value from any AI tool, your underlying data must be in order.
“Data readiness is an organizational leadership issue more than a technology problem.”
Q: That sounds like it requires significant organizational discipline before you even open a vendor conversation.
Gaertner: Exactly, and that distinction matters enormously. Data readiness is an organizational leadership issue more than a technology problem. It requires attention and investment at the leadership level. It can’t be solved by submitting IT tickets.
For a legal function spanning multiple jurisdictions, practice areas, and decades of legacy agreements, getting the data house in order is substantial work. But it is the necessary prerequisite, whether you have 1,500 contracts or 150,000. The good news is that there are AI tools that can streamline this effort by generating playbooks from current contract templates, and that can crawl across the enterprise to locate your signed contracts. However, you tackle this challenge, solving it is crucial. Skipping this critical step means you will be retrofitting from day one.
Q: Where does process alignment fit into this? Is data the only prerequisite, or are there others?
Gaertner: Process alignment is equally important.
AI agents thrive on repetitive, well-defined workflows. For contract review to work at scale, for example, you need playbooks, clause libraries, and agreement within the team on what “good” looks like before you ask an AI to help you get there. Even with the good news that some AI tools can help you build those playbooks, this only works effectively if you've already done the intellectual work of defining your standards.
This capability to automate tedious and repetitive, boring processes will be a significant win for our teams; they will love this as it allows them to focus on higher-value, more engaging legal work.
AI agents thrive on repetitive, well-defined workflows.
Strategy before selection
Q: What advice would you give a general counsel who's under pressure to “do something with AI” quickly?
Gaertner: Resist the pressure to move fast without thinking first.
The single most important thing not just a legal but any leader can do right now is define the questions before selecting the tools.
- What are we actually trying to accomplish?
- What tasks are creating the most friction?
- Where are we losing time that we should be reinvesting?
If you can answer these questions clearly, the tool selection becomes much more tractable. If you can't, no tool will save you.
In talent acquisition, it may even be worth considering: we need a standard question for all job interviews: "How do you use AI, which tools, which tasks, and how often?" to gauge current capabilities and attitudes.
Q: Is there a risk of being too cautious, though? Legal functions that wait too long may find themselves behind.
Gaertner: The risk of inaction is real, and I want to be direct about it.
Our outside counsel are using AI. Our vendors are. Our competitors are. Staying behind is not an option. Our team at ABB has a unique advantage, we are working for a technology company, and we are surrounded by engineers who live and breathe AI.
Staying behind is not an option.
“The goal is to be fast and deliberate. Staying behind is not an option.”
Q: How do you think about the competitive dimension, not just for the legal function, but for the lawyers themselves?
Gaertner: AI is not going to replace lawyers. But lawyers who use AI effectively will outperform those who don't, and over time, clients will notice.
The current moment is comparable in scope to the introduction of the internet, a shift so fundamental that those who engage thoughtfully now will reshape what their function is capable of, while those who wait will find themselves running to catch up.
I am optimistic about what this means for the profession. There will be no shortage of legal work. But the nature of that work will change, and the lawyers who thrive will be the ones who lead that change rather than resist it.
Furthermore, considering that all big law firms at Legal Week are using AI, we need to insist to get our share in their savings from these efficiencies.
Maintaining credibility and competence in the AI Era
Q: Where does governance fit into all of this? Is that a topic for legal or for IT?
Gaertner: It directly impacts our credibility as business partners.
If we are trying to advise our clients on AI without actively engaging with these tools ourselves, we risk losing that credibility. Questions about AI use involve what we represent to clients and regulators when we employ AI-assisted processes. They involve how we fulfill our duty of competence and supervision.
They involve what happens to privilege when documents are processed through a vendor's infrastructure. These are legal questions that require us to be actively engaged, not just informed after the fact. It's also important to understand the practical limitations; for example, AI can get “bored” when reading a long document, which necessitates proper human oversight and quality control.
We all need to embrace AI to understand its nuances and implications.
Q: Final question: What is the most important thing a legal leader should do this year?
Gaertner: Define where you are.
Before you run toward AI, understand your starting point, the state of your data, the maturity of your processes, and the sophistication of your team. Set a baseline. Then make a plan that is sequenced and realistic. The organizations that will succeed are not the ones with the most ambitious AI vision. They are the ones that execute deliberately, measure honestly, and build from a foundation that was actually designed to hold weight.
Implementing with patient ambition
The conversation with Mathias left me reflecting on what patient ambition looks like in practice. While it might be easy to assume that it is fear-based, this is an incorrect assumption. This approach is rooted in the discipline that comes from knowing exactly how much is at stake.
For legal functions grappling with the pace of AI change, his framework offers something valuable: a way to move with purpose rather than pressure, grounded in pragmatic observations and strategic considerations.
In the next installment of this series, we continue the conversation with another senior colleague whose perspective on the AI vendor landscape and outside counsel dynamics adds a different dimension to the picture.
Disclaimer: The information in any resource in this website should not be construed as legal advice or as a legal opinion on specific facts, and should not be considered representing the views of its authors, its authors’ employers, its sponsors, and/or ACC. These resources are not intended as a definitive statement on the subject addressed. Rather, they are intended to serve as a tool providing practical guidance and references for the busy in-house practitioner and other readers.
