Your firm is already running on AI. Most managing partners have no idea how much.
That is the argument Kathy Serenko, Amy Adams, and I made during our June 3rd webinar on “Three AI Conversations Every Law Firm Needs to Have”, hosted by the Law Firm Profitability Group on LinkedIn. Not which tools to use, but what your firm needs to have decided before it touches any of them. The recording is here.
Kathy Serenko is the founder of AI Efficiency Labs, and Amy Adams is the founder of Gaia Allies and AIReady™.
Kathy led with the governance perspective. Her opening question: Are you already exposed? Her answer, based on what she sees in firms daily: almost certainly yes.
AI enters your firm in ways leadership cannot see. A paralegal uploads 300 pages of medical records to a personal ChatGPT account under a deadline. The summary looks clean. Three months later, the error surfaces in a settlement discussion. A vendor tool your firm approved last year has AI inside it you never examined. The tools you signed off on produce errors that reach binding deliverables before anyone reviews them. AI is designed to produce plausible output, and it will not tell you when it is filling in blanks.
Kathy’s second point: having an AI policy is not the same as having governance. Less than 40% of law firms have a policy at all, and a policy is only a statement of intent. The missing layer is the day-to-day controls built into your actual workflow. Without those controls, the policy and the reality of what is happening in your firm will not align.
Governance means a domain expert, not IT, is the authority over AI output in each practice area. Someone who reads a research memo and recognizes when the content is wrong. The errors appearing in court sanctions are content errors, and the people catching them are judges and opposing counsel. They are domain experts. Your oversight authority in each practice area needs to be one too.
Amy made the workflow case. Her rule before recommending any AI system: map the workflow first. She recently spent 90 minutes with a family law partner doing exactly that. In your firm, that work probably has not been done. If the process exists only in someone’s head, the AI system inherits every gap in it. The tool scales what is already there.
Before the workflow is designed, the economics need to be settled. If a matter that used to take 10 hours now takes five, and your firm is still billing hourly, you may be cutting your revenue in half for that work. Pricing needs to be decided before the workflow is built. Firms that build the workflow first and sort out the pricing later often discover they cannot make money at it.
Everyone in your firm is asking the same question in different ways: what happens to me? The concern is real at every level, and dismissing it does not make it go away. What is actually changing is what each role does. Paralegals are running AI systems that used to require associate hours. Associates who used to produce first drafts are now verifying AI output and building judgment through that review. Partners who used to do the work are moving toward designing the systems that do it. The firms handling this well are having that conversation before the tools arrive.
Leadership does not get to delegate this responsibility. You cannot hand AI governance to a technology committee and check back in six months. You have to understand where the exposure is and whether your compensation structure supports what you are asking of your partners.
Your partners who put their expertise into an AI system will watch their billable hours fall as the system takes on that work. That contribution needs to be reflected in how they are compensated. If your firm still pays only for billable hours, you will not get the cooperation needed to build something that works.
The firms moving ahead well on this have answered the harder questions before reaching for tools.
