What I have learned about how lawyers actually change the way they work.
Highlights
- Lawyers change how they work when they see their own work done demonstrably better, not through feature lists.
- Trust grows when expert guides show the technology candidly and respond well to skeptical lawyers who find the gaps.
- Adoption spreads through colleagues, and AI-assisted workflows are giving under-resourced legal organizations back hours for bigger work.
A few months ago, one of our Solutions Engineers arrived for office hours at a law firm that had just gained access to CoCounsel Legal. Our contact at the firm offered a friendly warning: the lawyers were very busy, rarely attended software trainings, and almost no one had accepted the invitation.
Two lawyers came. Neither had ever logged into the product. After showing them how to sign in, our Solutions Engineer walked them through a task familiar to any litigator: reviewing a set of documents to identify deposition exhibits, then drafting an outline. Partway through, one of the lawyers interrupted her. “Hold on!” he said and called a colleague into the room so she could see it too.
By the end of the session, the audience of two had grown to forty. Word had traveled.
I return to that story often because it captures something I have come to believe about change in the legal profession. Technology is rarely the limiting factor anymore. Imagination is. Lawyers change how they work when they see their own work done differently, and demonstrably better than the way they do it now. I say this as a former litigator who skipped software trainings myself, for two reasons: I was busy, and I doubted that any tool could materially improve what I did.
For AI to improve legal work, then, it is not enough that the tool performs well. Lawyers must be able to picture the use cases, and the person showing them must understand legal work from the inside.
From where I sit
I began my career as a litigator and spent the better part of a decade at an AmLaw 20 firm. I later joined Casetext as an early member of the team that built CoCounsel, and I now lead Solutions Engineering for Legal Professionals at Thomson Reuters. That path has given me three vantage points on AI in legal practice: the courtroom, a startup, and one of the profession’s longest-established institutions.
What strikes me most is that the change is not only in what we build, but in who we place in front of customers. My team includes former practicing lawyers, AI solutions architects, engineers, and product specialists. Their role is not to deliver a standard demonstration. It is to sit beside a lawyer, understand how a particular piece of work gets done, and help redesign it.
The lawyer who asked, “What would I do differently?”
One of my favorite examples involves a criminal defense attorney preparing for a retrial. He had served as second chair in the first trial; this time the court had appointed him first chair. Working with one of our Customer Success colleagues, he uploaded the transcripts from the first trial and asked CoCounsel to critique each phase of his performance (his opening, each cross-examination, and his closing) against a relevant appellate decision. He then asked it to help him rebuild his opening statement.
What I find notable is not the speed, but how he chose to use it. He used the technology to examine his own performance, identify its weaknesses, and prepare more thoroughly for his client. The judgment remained his; the tool helped him sharpen it. This is what we mean at Thomson Reuters by “Fiduciary-Grade AI™”: technology built for professionals whose decisions carry real consequences.
The practice group that pushed back hardest
The best lawyers do not simply use AI; they hold it to their own standards.
Our team recently completed a five-part series of advanced clinics with a family law practice group. Few groups I have worked with have examined the technology as rigorously. They were skeptical, appropriately so, and they tested CoCounsel hard. By default, CoCounsel links to the authorities it consulted on Westlaw, Practical Law, and other sources in reaching its conclusions. The group wanted more. They asked it to explain, for each authority cited, why that authority was relevant.
By the final session, the group had become comfortable using the verification features to follow the reasoning behind CoCounsel’s research, and they were ready for more advanced work. One associate demonstrated a workflow she had built herself: using CoCounsel to check and prepare child support calculation worksheets, both simple and complex. This work is stressful and error-prone when it depends on the state’s calculator alone. The practice group chair then named the leaders of two other practice groups she wanted us to meet, asked us to return for recurring clinics every six months, and told us that her colleagues “don’t realize what they’re missing.”
The lawyers who pushed back hardest became the strongest advocates, which I do not think is a coincidence. Trust is not built by a flawless demonstration. It is built by how we respond when a lawyer finds the gap.
Expediting criminal justice reform
Our AI for Justice program has shown me how far this work can reach. One innocence organization we support is studying eyewitness identification policies, since eyewitness misidentification is among the leading causes of wrongful convictions. The first version of their study (without CoCounsel Legal) took four years. This time, they are examining roughly 450 policies across California, most of which fall short of legal requirements. With our team’s help, they built an AI-assisted workflow in CoCounsel Legal that captures more than 100 data points from each policy, and they now expect to finish in just a couple of months.
For under-resourced organizations, structural reform is often a wish-list item because there are simply not enough hours. When technology returns some of those hours, the work they can take on begins to change.
What we have learned
Across hundreds of law firms, courts and other government bodies, and legal aid organizations, three lessons keep recurring.
Start with the work, not the tool. Lawyers do not adopt features; they adopt better ways of doing the work in front of them.
Expertise and candor earn trust. When the person showing the technology has also drafted summary judgment motions, closed deals, or sat first chair at a trial, lawyers engage more readily, and more so when we are frank about what the technology cannot yet do.
Change spreads through colleagues. Every “Hold on!” moment gives the next lawyer a reason to try. Every practice group chair who says “you should talk to my colleague” opens another door.
Chosen by changemakers
When I consider what it means to be chosen by changemakers, I think of the defense attorney who asked what he could do better, the family law group that would not accept an answer on faith, the innocence organization rethinking years of work, and the lawyers who filled a training session that almost no one had planned to attend. What defines them is not how quickly they adopt new technology, but their willingness to ask how their work could be better and then act on the answer.
Helping lawyers answer that question is my team’s job and our privilege: one workflow, one practice group, and one “Hold on!” at a time.
The best don't just do their work. They change what is possible.
The right technology can change everything
