Nobody Wants More AI Chat: How We Built AI That Grew Our Revenue 78% YoY and Performs Real Work Across HR, IT, and Finance with Rippling

Most AI in workforce software still stops at the chat window. It can answer questions, summarize data, and sound helpful — but it rarely prepares real operational work for execution. In this session, we’ll show how we built Rippling AI to stage real work across HR, IT, and Finance. Through a live walkthrough, attendees will see how an admin starts with a business question, gets an exact answer grounded in real company data, and turns that result into a staged action for review and approval — whether that’s updating payroll inputs, changing system access, or making org changes. Along the way, we’ll also unpack the AI concepts behind the experience: what it means to ground answers in a system like Rippling’s Employee Graph, how permissions and context shape trustworthy outputs, where automation should stop and human review should begin, and why moving from “chat” to “action” is the real design challenge for enterprise AI. We’ll share the product and design choices that made this possible, including inheriting permissions automatically and keeping humans in the loop before anything sensitive goes live. Attendees will leave with both a practical understanding of how operational AI systems work and a framework for building AI workflows that people can actually trust and delegate work to — not just chat with.