
Episode #413
Agentic AI in CPQ and Quote-to-Cash: Where Autonomy Belongs | Sourabh Kukar, Salesforce
Copilot, assistant, agent — the terms get thrown around interchangeably, but they describe three very different levels of AI autonomy. So where does agentic AI actually belong in the configure-price-quote process, and where should humans stay firmly in control? In this episode, Frank sits down with Sourabh Kukar, Director of Technical Consulting at Salesforce, where he leads a team of enterprise architects working across Revenue Cloud, Agentforce Revenue Management, CPQ, and H&A solutions for large enterprise customers. Sourabh was named Salesforce's FY24 CMT Technical Architect of the Year, holds 23 industry certifications, and is an IEEE Senior Member. Sourabh brings a rare mix of backgrounds to the conversation — starting as a Java developer, spending a decade as an integration architect, then moving deep into the Salesforce platform in 2019. That combination of coding roots, systems breadth, and platform depth shapes how he thinks about which parts of quote-to-cash are ready for real AI autonomy and which still need deterministic guardrails and human sign-off. Frank and Sourabh also dig into how to architect multiple specialized agents so one agent's bad output doesn't become another agent's bad input, why fixing your data and simplifying your processes matters more than the AI layer itself, and what changes for revenue operations teams as quoting and approvals move into conversational tools like Slack. In this conversation we cover: The real difference between a copilot, an assistant, and an agent — and where Salesforce's Agentforce quoting agent fits Which parts of configure-price-quote are ready for high AI autonomy, and where humans need to stay in the loop How to architect multiple specialized agents so errors don't cascade downstream Why fixing your data and simplifying your processes matters more than the AI layer on top What changes architecturally when quoting and approvals move into Slack A practical framework for choosing phased coexistence over a full CPQ-to-Revenue-Cloud migration If you're a Salesforce CPQ customer evaluating a move to Revenue Cloud, an enterprise architect thinking through agentic AI governance, or a revenue operations leader trying to separate real AI capability from hype, this episode offers grounded, practitioner-level insight from someone building these systems for enterprise customers every day. Learn more: Sourabh Kukar on LinkedIn Don't forget to subscribe on YouTube and your favorite podcast platform. More CPQ research and resources at https://novuscpq.com






