When our own leaders publish on where AI is headed, we like to share the thinking behind it. In an article for Law.com‘s Legaltech News, our CTO John Wei and colleague Animesh Kumar, SVP of contract, compliance and commercial services, dig into a change that legal, compliance, and technology leaders are all facing together.
AI has grown up. It is no longer just chatbots that answer questions. It is agents that plan work, use tools, and carry out tasks with barely any oversight. That is powerful, but it comes with a catch. The more freedom an agent has to act on its own, the smaller the gap between a minor slip and a real problem for the business.
How good your AI is depends on how you govern it, not on which model you happened to pick.
They explain a common mistake they call AI sprawl. When every team grabs its own tools, no one can say for certain which agent has access to which data. You cannot keep watch over what you cannot see. Their fix is to run AI through one central platform so leaders get a single clear view and can apply the same rules to every agent at once.
They also argue against inventing brand new rules for AI, pointing instead to trusted security standards that already exist. One example lands hard: an agent reading an outside document could run into a hidden instruction buried in the text, and a poorly governed agent might just follow it.
The risk they stress most is what happens when an agent has big goals but loose limits. Picture an agent running a contract library that finds a one-time exception for a single client and then wrongly applies it across every contract. A single decision quietly becomes company-wide policy. The answer is to give each agent a tight, clearly defined role, and to keep high-stakes moves under human sign-off.
Their takeaway is that keeping people in the loop does not slow AI down. It is exactly what lets a business move fast without losing control. It is the kind of practical, forward-looking thinking our team brings to the way we help clients adopt AI.
John and Animesh lay out the full set of steps in their article for Law.com: Why AI Strategy Fails Without Governance, And What Legal and Compliance Teams Can Do About It
