Tim Shea, Founder and CEO of Latticework Insights, joined the Transform NOW series (from SS&C Blue Prism) to talk about a question most companies skip past on their way to buying an agentic AI tool: what actually has to change about how a team is structured for an AI agent to be trustworthy enough to put into production?
Standing up an agent is the easy part. The harder, and far less talked-about, part is organization design — who owns an agent’s decisions, how it gets evaluated before it ever touches a real customer or a real dollar, and who’s accountable when it’s wrong. Tim draws on Latticework’s own hands-on experience building agentic systems to lay out what a company’s structure needs to look like before “agentic AI” is a safe thing to say out loud to a board, rather than a plan that quietly collapses the first time an agent makes an expensive mistake.
If your team is standing up its first AI agent and isn’t sure who should own it once it’s live, get in touch — that’s exactly the kind of organization-design problem Latticework helps sort out.
