What does it take to think like a product leader in the AI era? In this episode of our virtual speaker series, Salesforce Senior Product Lead Vasanthi Neelagiri speaks on mastering the AI PM mindset. This session examines how product leaders can shift from shipping features to driving real outcomes, focusing on value creation rather than model deployment. It expands the product toolkit with AI-specific frameworks that build on established practices, from data readiness to experimentation to responsible design. It also highlights the role PMs play as stewards of trust, grounding innovation in fairness and clarity. Participants will walk away with practical mental models they can put to work immediately—how to evaluate whether AI is the right solution, how to design effective data and product loops, and how to clearly communicate both risks and value to stakeholders.
Join us for new conversations with leading product executives every week. Roll through the highlights of this week’s event below, then head on over to our Events page to see which product leaders will be joining us next week.
Show Notes:
- AI is not replacing Product Managers (PMs); rather, it elevates their role by allowing them to focus on strategy, empathy, and judgment.
- AI automates low-value, repetitive PM tasks (like note synthesis and basic reporting) and enhances high-value tasks such as strategic thinking and prototyping.
- AI products are “probabilistic,” meaning they deliver suggestions with confidence scores and inherently carry uncertainty—requiring new methods to manage trust and truth.
- The evolution of PM in the AI era is structured around three pillars: personal productivity (using AI tools), building real AI products (focusing on probabilistic systems), and transforming organizations (AI-powered internal tools).
- Avoid “AI theater”—using expensive AI for simple rule-based problems is inefficient and risky.
- Use AI where its probabilistic and inferential abilities bring clear value (e.g., for prediction, classification, clustering, and creative generation).
- Success with AI products depends on defining and monitoring clear, measurable, and relevant success metrics (product KPIs and model metrics).
- Always have fallback (“human-in-the-loop”) strategies—particularly for high-stakes scenarios like finance or healthcare—to ensure user control and safety.
- AI product development requires a shift from linear build-test-ship processes to continuous monitoring, retraining, and model evaluation.
- Proper AI testing involves multiple layers: offline evals on historical data, shadow testing in live environments, and final user-facing A/B tests.
- Model drift and data shifts must be actively monitored; AI models need maintenance even if no code changes.
- Define “dual track” metrics: both user/product KPIs (e.g., adoption, conversion, NPS) and technical model metrics (e.g., drift, latency, hallucination rate).
- Responsible AI requires addressing four pillars: fairness/bias mitigation, transparency/explainability, privacy/security, and accountability.
- PMs play a key role in translating high-level ethical guidelines into actionable product requirements.
- McKinsey’s four skills for AI PMs: risk management/compliance, empathy/trust-building, low-code prototyping, and agentic (AI workflow) planning.
- Ethical failures in AI are more damaging than functional bugs, potentially causing reputational and regulatory crises.
- For executives, communicate the strategic value, risks, and proprietary advantages of AI investment.
- For customers, build trust through transparency (e.g., “AI-generated label”), human control, and clear user experiences.
- For cross-functional teams, champion data quality and enforce accountability for AI outcomes and failures.
- Numerous AI-powered tools exist to help PMs with prototyping (Builder.io, Figma), PRD drafting (Notion AI, Gemini), structure visualization (Miro, Whimsical), and meeting summarization (Otter.ai, Fireflies.ai), enabling faster, more autonomous product development.