How do product leaders decide what AI should automate and what should stay human-in-the-loop? In this podcast hosted by Mphasis VP of Products Chenny Solaiyappan, Level AI Head of Product Sirisha Machiraju will be speaking on building trusted agentic AI for customer experience. She also shares a practical playbook for evaluating assistive vs. agentic workflows and explains how conversation intelligence is reshaping customer experience across industries, including healthcare.
Subscribe to the Product Talk podcast on Spotify and Apple Podcasts and catch every conversation with leading product executives.
Show Notes
- Product managers create alignment by turning ambiguity into structure and momentum across teams.
- Strong product leadership often starts with cross-functional influence before formal leadership roles emerge.
- AI becomes strategic when it shifts from experimentation to being embedded into core workflows and decision-making.
- Organizations like Uber treated machine learning as foundational long before generative AI became mainstream.
- Data-driven experimentation should be a default operating model, not a special initiative.
- Internal adoption of AI tools is as important as customer-facing adoption.
- Conversation intelligence platforms unlock hidden business signals inside customer interactions.
- Customer sentiment data can directly influence product strategy and roadmap prioritization.
- AI enables companies to transform millions of support interactions into structured insights.
- Agentic AI introduces new opportunities to rebalance work between humans and automation.
- Choosing between assistive AI and full automation depends on risk tolerance, reversibility, and reproducibility.
- High-risk industries like healthcare require stronger guardrails before deploying autonomous workflows.
- AI rollout should follow a maturity curve rather than a binary automate-or-not decision.
- Reversibility is a key design principle for production AI decisions.
- Reproducibility strengthens confidence when relying on non-deterministic LLM outputs.
- Trust—not capability—is the biggest barrier to AI adoption in enterprise environments.
- Adoption requires partnership with customers, not just shipping technology.
- Product teams increasingly operate as “product builders,” using AI tools to prototype faster and reduce documentation overhead.
- Visual artifacts and interactive prototypes are replacing long written PRDs in modern product workflows.
- Agentic AI will likely transform everyday coordination tasks—like scheduling services or navigating insurance complexity—into seamless automated experiences.
About the speaker
Sirisha Machiraju is a product leader with 15+ years of experience of product development. She has held leadership roles at Uber, Adobe, and Yahoo. At Uber, she was instrumental in building the platform that enables agentic AI contact resolution across digital channels optimizing cost to serve and increasing customer CSAT. Deeply customer-centric, Sirisha excels at aligning teams around a clear product vision and strategy and bias for action. She also has experiences in domains of MarTech, supply chain, data platforms. She is currently the VP of product at Level AI (https://thelevel.ai/).