PMMs are getting forced into a bad tradeoff on the marketing site. Old-school chatbots can push visitors to a CTA, but they fall apart on real, nuanced questions. New AI SDRs can answer almost anything, but they often wander and don’t reliably move people to the right next step. This session gives a practical framework to get both: match intent to the right CTA (lead magnet, email follow-up, meeting, or self-serve signup) and add enough control and visibility to steer generative AI toward measurable conversions—without going back to rigid decision trees. Key Takeaways: - The intent to CTA playbook: route common intents to the right next step so chat doesn’t force “book a demo” on everyone. - Control without a rigid decision tree: how to steer generative AI (offers, claims, tone, handoffs) while keeping deep, contextual answers. - What to measure: the few leading indicators PMM can actually move (offer shown rate, path completion, handoff quality) to drive conversion gains.