Start Small, Build Together: A Practical AI Path for Law Firms
Law firms do not need to have everything figured out before they begin. That was Dan Paquette’s most useful takeaway from this year’s Power Platform Community Conference in Las Vegas, Nevada, where he spent much of his time exploring Copilot development and governance. The lesson was not a particular feature or a new way to build an agent, but a more practical idea about getting started.
The Power Platform Community Conference brought together people at very different stages. Some have deployed agents across large organizations. Others were building their first skill. Experienced builders shared what worked, what failed, and what they would change. That willingness to learn in the open made the distance between getting started and making meaningful progress feel much smaller.
For law firms still early in AI adoption, that willingness to share lessons is encouraging. A practical path forward is to start with a manageable problem, learn from the people doing the work, and build governance alongside the solution rather than adding it afterward. Firms do not need to wait for a complete AI strategy or chase the largest possible transformation to begin.
More People Can Solve Their Own Problems
Building a useful solution is no longer exclusively the work of someone who can write code. For Dan, that shift highlighted how approachable the tools are becoming. As skills, agents, and workflows mature, non-technical people have more ways to turn their knowledge of a task into something that helps them do it. They can begin building themselves instead of handing every idea to a technologist and waiting for an application or automation.
That does not make technologists unnecessary. It changes where they can contribute most: providing a safe environment, helping people choose the right tool, and supporting integrations or complexity beyond a beginner’s reach. The person closest to the work brings the process knowledge; technology teams help make the solution reliable and appropriate for the firm. That is a partnership, not a handoff.
Start with the Work, Not the Agent
With that partnership in place, the next question is where to begin. Look closely at the processes in your own job: What do you repeat every day? Where do you copy information between systems? Which requests follow the same steps? Where do you spend time finding information instead of using it? Those questions are a better starting point than asking how many agents the firm should build. AI is not about to take over the world or make every part of a law firm run itself. But it is becoming useful enough to make that closer look worthwhile.
The goal is to choose the lightest tool that solves the problem, not to use AI simply because it is available. In some cases, AI might help a person complete a task more easily. In others, a conventional workflow might automate it entirely. Possible starting points include helping an assistant prepare a routine internal request, locating an approved policy, or moving a predictable administrative task through a review process. These are illustrative examples, not a prescription for every firm.
Build Guardrails into the First Experiment
Choosing the lightest tool is only part of designing the solution. Making the tools accessible does not mean opening the doors without boundaries: in a law firm, confidentiality, access permissions, and professional judgment need to be part of the design from the beginning. Even a small experiment should have a clear owner, approved data sources, an understood audience, and a way to check whether the output is useful and accurate. Where a mistake could affect a client or a legal matter, human review should remain part of the process.
Governance sessions at the conference reinforced the questions behind those guardrails: who can build, what can an agent access, and who takes responsibility when its builder moves on? These are not reasons to avoid experimenting. They are reasons to give people a supported place to begin. Start with a low-risk task and approved information, test with a small group, share the results, and adjust before expanding.
Progress Does Not Have to Start Big
Dan’s takeaway from Las Vegas is that the opportunity for law firms is not just better technology. It is more people being able to improve the work they know best. Give them room to start small, support them with practical guardrails, and encourage them to share what they learn. A modest solution that makes a real task easier is a worthwhile beginning… and a better foundation for what comes next than waiting until every question has an answer.
Stay Connected
Looking for more ways to interact with Kraft Kennedy?
Check out our upcoming events!