Agentic AI Execution

Leena Agarwal

The podcast for TPMs, product managers, and builders shipping real AI systems. No hype, no frameworks — just honest conversations about what it actually takes to move from model to production. Topics include agentic AI, AI program execution, stakeholder alignment, technical roadmaps, and building trust in AI-driven orgs. Hosted by Leena Agarwal, 18-year TPM veteran from Amazon and Meta. New episodes every week. Built for tech leaders who ship.

Episodes

  1. 4 HR AGO

    What does a responsible agentic AI workflow actually look like?

    Ilya co-founded AWS Chatbot — now Amazon Q for Developers — before large language models even existed. Today he builds agentic AI platforms and has one of the most thoughtful personal AI systems I've come across. We worked together at AWS, and this conversation reminded me why his thinking has always been a step ahead.00:00 Introduction02:18 Why Ilya runs every AI workflow manually03:34 The local Git repo that acts as his personal AI assistant05:32 The /recap skill — daily email, Teams + transcript synthesis07:17 Voice memo pipeline using Whisper — zero data leaves his machine11:02 The Telegram bot that captures ideas while driving13:10 Bringing teams along on the AI journey18:07 Why demos beat mandates — the cultural shift21:20 AI as a trusted colleague — the Arnab analogy24:15 The mission beneath all this agentic work27:03 "It's not doing more with less. It's doing the same with less."🔗 Resources & Connections:Connect with Ilya: https://www.linkedin.com/in/ilyabezdelev/Connect with Leena Agarwal: https://www.linkedin.com/in/leenaagarwal/

    28 min

  2. 2 DAYS AGO

    Vibe Coding Won't Save You — Here's What Will | Agentic AI Execution Ep. 5

    Former Amazon Senior PM and Maven instructor GK joins the show to explain why "vibing" with AI isn't enough to build professional-grade systems. She introduces Spec Engineering—her tactical framework for moving beyond the prompt to build high-quality, deterministic AI agents.📌 Timestamps:0:00 - Introduction to GK2:15 - AI as a tool for Neurodivergence8:40 - Why Vibe Coding doesn't scale12:50 - The Spec Engineering Framework18:30 - Future of the PM workforce24:10 - Stop consuming, start thinking27:50 - Key Takeaways 🔗 Resources & Connections:Connect with GK: https://www.linkedin.com/in/gayathri-keerthana/Connect with Leena Agarwal: https://www.linkedin.com/in/leenaagarwal/#AgenticAI #ProductManagement #SpecEngineering #AIPM #VibeCoding #ArtificialIntelligence

    30 min

  3. 3 DAYS AGO

    AI Agent Signed a Legal Contract in Vietnam (Without My Permission) 📉 | Agentic AI Execution Ep. 4

    In 2026, 36% of all new ventures are solo-founded. We are witnessing the birth of the "Invisible Enterprise"—a model where human founders stop "doing" and start "directing" a swarm of autonomous agents to handle everything from revenue operations to legal negotiations.In this episode of Agentic AI Execution, I sit down with Tim Cortinovis to discuss the shift from "Vibe Coding" prototypes to production-grade Agentic Revenue Systems. Tim shares the wild (and legally eye-opening) story of "Rick," an AI agent that autonomously negotiated and signed a publishing contract in Vietnam before Tim could even review the terms.Whether you’re a founder, a Technical Program Manager, or a leader navigating the AI transition, this is your blueprint for achieving massive leverage in the agentic era.In this episode, we cover:The "Rick" Incident: What happens when an agent handles legal affairs without human intervention?Solo-Unicorns: The data behind the 2026 trend where one person runs what used to take a team of 20.The Leverage Shift: Moving from "Doing" the work to "Directing" the agentic swarm.Revenue Systems: Why your CRM data is the silent killer of your AI strategy.📌 Timestamps:0:00 – The AI agent that signed a Vietnamese contract.4:15 – What exactly is an "Invisible Enterprise"?12:40 – Scaling to $1M ARR with zero employees.22:10 – The future of the "Wartime AI CEO."🔗 Resources & Connections:Connect with Tim Cortinovis: https://www.linkedin.com/in/timcortinovis/Connect with Leena Agarwal: https://www.linkedin.com/in/leenaagarwal/#AgenticAI #SoloUnicorn #InvisibleEnterprise #AILeadership #VibeCoding #RevenueOperations #AgenticROI #StartupScale #AIGovernance

    25 min

  4. 27 APR

    Why 95% of AI Agents Never Ship (and How to Be the 5%) | Agentic AI Execution Ep. 3

    95% of AI agents never leave the demo stage. They look great in a prototype — and then collapse under real production pressure.In this episode, Jyothi Nookula (ex-AWS, Meta, Netflix | 12 patents | 1,500+ PMs coached) breaks down exactly what separates the agents that ship from the ones that don't — and what it means for product managers building in the agentic era right now.What you'll learn:→ Why the MIT-cited 95% failure rate happens — and the specific production gaps killing your pipeline→ How the PM role has fundamentally changed: you now own evals, prototypes, and golden datasets — not just PRDs→ The "Ferrari vs. Camry" framework for knowing when AI is the wrong tool entirely→ Why human-in-the-loop is a power design pattern, not a fallback→ How to manage leadership expectations when your system is probabilistic but they're thinking deterministically→ Practical tips for building with Claude, MCPs, and skills without blowing your context windowIf you're a PM, TPM, or engineering leader trying to move from AI demo to production reality — this one's for you.⏱️ Chapters0:00 – The 95% Demo Gap: Why agents fail before they ship02:20 – Hardening agents: Why this isn't like traditional dev05:12 – What AI PMs actually own now (hint: it's not just the PRD)08:30 – Evals deep dive: Happy paths, edge cases, and prompt injection11:00 – Prototyping with Claude, MCPs, and keeping context lean15:30 – The Ferrari & the Camry: When NOT to use AI20:00 – Closing the leadership expectation gap25:00 – Agentic pipelines vs. traditional pipelines31:00 – Human-in-the-loop as a trust architecture🔗 ResourcesNextGen Product Manager → https://nextgenproductmanager.com/Connect with Jyothi → https://www.linkedin.com/in/jyothinookula/Connect with Leena → https://www.linkedin.com/in/leenaagarwal/#AgenticAI #ProductManagement #AIEngineering #AIPM #MachineLearning #LLM #TechPodcast #ProductStrategy #AILeadership #ClaudeAI

    27 min

  5. 21 APR

    AI Is Killing Junior Devs. A 23-Year Amazon Vet Tells The Truth | Agentic AI Execution Ep. 2

    Is AI quietly ending the junior developer career path?Leah Melvoin spent 23 years at Amazon — she joined when the company had ~100 employees and left as a Principal TPM who personally resolved over 200 engineering pain points. She created the Voice of the Developer program and wrote the leveling guides thousands of engineers still use today.In this episode, she gives an unfiltered take on what agentic AI is actually doing to engineering careers — and what developers at every level should do right now.We cover:→ Why junior dev placement rates are at 6–12% and what that means→ The 2-pizza team model — why it finally works in the agentic AI era→ How AI is creating a dangerous skill atrophy trap for engineers→ Why senior engineers are the most secure (and what mid-level devs must do)→ What leaders are getting dangerously wrong about AI-driven layoffs→ The one career move Leah recommends if you're just starting outThis is not hype. This is 23 years of pattern recognition applied to a moment that's changing fast. ——🎙 Agentic AI Execution — real stories from people shipping AI in the real world Connect with Leah Melvoin: https://leahmelvoin.com CHAPTERS 0:00 Intro — Who is Leah Melvoin? 1:08 What was Voice of the Developer at Amazon? 3:41 What startups can learn from Amazon's engineering culture 4:37 The 2-pizza team model is back — and it's 4 people 13:20 Junior vs senior engineers in the agentic AI era 15:19 What leaders are getting wrong about AI and layoffs 18:45 Why coding is no longer the most important skill 20:23 The skill atrophy trap — are devs losing their ability to code? 23:10 AI productivity paradox — 90% usage, 10% gains? 29:35 One piece of advice to survive the agentic AI era#AgenticAI #SoftwareEngineering #TechCareers #AIJobs #EngineeringCulture #Amazon #TechLeadership #JuniorDeveloper #FutureOfWork #Podcast

    33 min

  6. 13 APR

    Former Meta TPM: Why Most AI Programs Fail | Jenny Truong

    Most AI programs fail—not because of the technology, but because of broken trust and misaligned decision ownership. In this debut episode of Agentic AI Execution, Jenny Truong (14 years at Meta leading Newsfeed and Stories) joins Leena Agarwal to break down the honest, practical playbook for executing AI programs that actually scale. In this episode, we cover: ✦ The APP Framework: Acknowledge, Plan, and Pivot through AI shifts. ✦ Trust as a Metric: Why 95% of AI initiatives fail and how to fix it. ✦ Signal over Noise: Distinguishing between data signals and people signals. ✦ The Evolving TPM: Why critical judgment is the skill AI can never replace. ✦ Leadership: Navigating high-pressure stakeholders and being a voice in the room. Whether you're a TPM, PM, or tech leader, this is your guide to navigating the complexities of AI execution in 2026. #TPM #TechnicalProgramManagement #AIExecution #AgenticAI #WomenInTech #AILeadership #MetaAlumni #TechPodcast #ProductManagement

    27 min
  • 6 Episodes

About

The podcast for TPMs, product managers, and builders shipping real AI systems. No hype, no frameworks — just honest conversations about what it actually takes to move from model to production. Topics include agentic AI, AI program execution, stakeholder alignment, technical roadmaps, and building trust in AI-driven orgs. Hosted by Leena Agarwal, 18-year TPM veteran from Amazon and Meta. New episodes every week. Built for tech leaders who ship.

Information

  • Creator
    Leena Agarwal
  • Years Active
    2k
  • Episodes
    6
  • Rating
    Clean
  • Copyright
    © Leena Agarwal
  • Show Website
    Agentic AI Execution