• Lead conversation design for Salesforce’s website AI agent, improving clarity, tone, and task completion across sales, product exploration, and support-adjacent intents.
• Developed escalation flows that increased qualified sales handoffs and reduced support-seeking dead ends by grounding prompts in user intent rather than generic call-to-action triggers.
• Created frameworks for intent recognition, routing logic, and moment-level scoring to better distinguish top-of-funnel curiosity from high-intent buying signals.
• Partnered with PMs, engineers, and data teams to refine instruction sets, guardrails, and behavior models across releases, driving measurable improvements in relevance and lead capture.
• Improved agent handling of ambiguous, conceptual, and comparison queries through updated taxonomy guidance and instruction rewrites.
• Designed conversation patterns that strengthened brand trust, reduced confusion, and aligned the agent’s voice with Salesforce’s product positioning.
• Introduced qualitative feedback loops and utterance-pattern analysis to identify adoption gaps, friction points, and optimization opportunities.
• Helped evolve the agent from a passive support persona to a proactive digital rep that guides users toward solutions, content, and sales pathways.
