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.

  1. 15 hr ago

    Builders vs. Marketers with AI | Episode 21

    As Preethy highlights in the podcast, technical founders frequently fall into the trap of thinking "if I build a great product, it will sell itself". But as engineering and AI code generation become commoditized, building is no longer the bottleneck—customer validation and distribution are. Timestamps / Chapters 00:00 - Introduction: Why Advanced AI Isn't Everything 00:19 - Preeti’s Journey: 25 Years in Silicon Valley 00:45 - Evaluating AI Claims: Privacy, Hallucinations, & Security 01:15 - Questions Founders Must Answer Before Shipping AI01:30 - Agentic AI Capabilities & B2B vs. B2C Dynamics 02:00 - 2026: Moving from AI Adoption to Enterprise Production 02:30 - Solve Unsolvable Problems: Guardrails & Build vs. Buy 03:00 - Blitzscaling Speed vs. Cost & Safety Trade-offs 03:30 - 10x Growth Strategies & Agentic AI in Marketing 04:00 - What Technical Founders Get Wrong About Marketing 04:30 - Human Connection & The 3S Formula for Success

  2. 6 days ago

    The Perception Gap: Why Great Work Gets Overlooked | Episode 20

    Are you doing the best work of your career, yet someone else keeps getting the credit? In this episode of Agentic AI Execution, host Leena sits down with Uma, a leadership coach and the author of Sought After Leader. After spending more than two decades at Microsoft, Uma realized a hard truth: simply being great at your job and executing flawlessly is no longer enough to get you noticed—especially as AI takes over the execution layer. Tune in to learn how to make your value undeniable and ensure your name is spoken in the rooms where decisions are made. 🕒 Timestamps:00:00 - Introduction: Why great work isn't enough 00:52 - Uma's journey from Microsoft to writing Sought After Leader 04:22 - The SOAR Framework explained 07:54 - The H.U.M.A.N. Edge: 5 skills to stand out in the AI era 13:05 - Real-world case study: Closing the perception gap 20:48 - Strategic visibility vs. bragging 23:42 - Finding "Level 10" opportunities in the white space 31:40 - Final thoughts and where to find Uma 33:57 - Episode recap and takeaways 🔗 Resources & Links:* Read the Book: https://www.amazon.com/Sought-After-Leader-Great-Enough-Above-ebook/dp/B0H7L5W4VM/* Connect with Uma: https://www.linkedin.com/in/uma-subramanian-limitless/* Connect with Leena:

  3. 4 Aug

    Will AI Take Your TPM Job? | LinkedIn Senior Staff TPM Ankur Gupta| Ep 19

    Could your program survive two weeks without you? If the answer is no, you’re not running an operation—you’re a single point of failure with a job title.In this episode of Agentic AI Execution, host Leena sits down with Ankur Gupta, Senior Staff TPM leading AI Infrastructure at LinkedIn (and former tech leader at Walmart). They tackle the burning question: Will AI replace technical program managers?Ankur breaks down why AI isn't taking your job, why status reporting was never the core value of a TPM, and how to build a "TPM Operating System" that runs seamlessly with AI agents while keeping human judgment at the center.---📌 KEY TAKEAWAYS & HIGHLIGHTS:• The Two-Week Test: How to structure programs that don't depend on your memory.• Status vs. Judgment: Give agents the status updates, but keep judgment to yourself.• Information Centralization & Work Decentralization: Scaling execution across 200+ teams.• The TPM Operating System: Operational, Collaboration, and Technical lenses.• Guardrails Over Memos: How to build real AI governance directly into technical systems.• Deterministic vs. Non-Deterministic Reliability: Managing AI agents that take different paths to a goal.---⏱️ TIMESTAMPS:00:00 - The Two-Week Test: Are You a Single Point of Failure?01:03 - Ankur Gupta's Career Journey to LinkedIn AI Infrastructure02:20 - What Agentic AI Execution Means in Practice03:30 - The Evolving Role of a TPM in the AI Era06:05 - Information Centralization vs. Work Decentralization08:45 - Moving from Coordination to Strategic Tech Leadership11:20 - The "TPM Operating System" Framework17:12 - Live Walkthrough: Set Once, Run Forever Checklist22:50 - What Should NEVER Be Handed Off to an AI Agent?25:10 - Lowering the "Cost of Yes" & Influencing Humans28:30 - Who Holds Accountability When AI Agents Ship Code?30:50 - AI Governance & System Guardrails34:10 - Reliability in Non-Deterministic AI Systems37:40 - What TPMs Need to Do in the Next 12 Months to Stay Relevant42:15 - Practical Hands-On AI Hacks for Daily Workflows46:10 - Overcoming AI Anxiety & Final Advice49:18 - Key Takeaways & Episode Wrap Up🔗 Connect on LinkedIn: Ankur : https://www.linkedin.com/in/agupta03/Leena: https://www.linkedin.com/in/leenaagarwal/

  4. 30 Jul

    Build Structure - Don't Chase Tools | Ep. 18

    In this episode, I sit down with Michael—a technical product management leader who has spent the last ten years building TPM teams from scratch. Together, we explore what agentic architecture looks like inside an engineering org and the mentorship that separates teams winning with AI from those who are just experimenting. Michael shares his mental models for AI adoption, explaining why we need to move past the "experimentation phase" and focus on building durable systems. Whether you are a product manager, TPM, or engineer, this conversation provides practical advice on shifting your perspective from using AI as a tool to leading AI as a teammate. ——————————————————————⏱ CHAPTERS——————————————————————00:00 - Introduction to Agentic AI Execution 00:31 - Michael's Journey & Building a TPM Team 06:07 - Why 95% of AI Agents Fail to Reach Production 15:51 - How Engineering Teams Can Close the AI Adoption Gap 12:54 - Leading AI as a Teammate, Not a Tool 23:19 - Outro & TPM Breakdowns Substack Links & ResourcesMichael's substack: https://michigoetz.substack.com/aboutConnect with Us! Michael: https://www.linkedin.com/in/michi-goetz Leena:https://www.linkedin.com/in/leenaagarwal

  5. 28 Jul

    Is Agentic AI FOMO Real? How to Lead When Half Your Team is AI Agents | Ep. 17

    In Episode #17 of the Agentic AI Execution Podcast, host Leena sits down with Myra Deshmukh—a 10-year Amazon veteran and founder of Leadership Lab—to break down what leadership actually looks like when half your team consists of AI agents. If you've been feeling "Agentic AI FOMO" or overwhelmed by how fast tech is moving, this episode is a must-watch. Myra shares her golden rule for AI, how to avoid outsourcing critical thinking, and why human judgment is your greatest asset in an automated world ---------------------------------------------------⏱️ TIMESTAMPS---------------------------------------------------00:00 - Intro: Leading when half your team is AI agents 00:32 - Welcome Myra Deshmukh (Leadership Lab & Ex-Amazon) 01:15 - Dealing with uncertainty & Agentic AI FOMO03:40 - Delegate work, keep judgment: Avoiding over-validation09:30 - How leadership style is changing in top tech companies 14:10 - Simple ways to build authentic relationships & trust at work 21:15 - The Manager's Insurance Policy: Verifying agent output29:50 - What to do if you feel overwhelmed or behind with AI 38:10 - Final advice for leaders & wrapping up ---------------------------------------------------🔗 RESOURCES & LINKS---------------------------------------------------• Check out Leadership Lab: https://www.leadership-lab.co/• Connect with Myra Deshmukh: https://www.linkedin.com/in/myra-deshmukh/• Leena | Agentic AI Execution Podcast https://www.linkedin.com/in/leenaagarwal/ #AgenticAI #Leadership #ArtificialIntelligence #FutureOfWork #ProductManagement #LeadershipLab #TechLeadership

  6. 22 Jul

    What actually separates teams that ship AI from teams that are just demoing it?

    With AI accelerating data flow, human alignment and judgment have become the ultimate bottlenecks in product development. If you are still relying on traditional, slow decision hierarchies, you are falling behind in the age of agentic intelligence.In this episode of *Agentic AI Execution*, Anjali—who leads AI programs on Google Maps and mentors over 100 startup founders—breaks down how to deliver AI products at a scale of billions. We explore why human-in-the-loop is still non-negotiable despite rising automation, how the Technical Program Manager (TPM) role is morphing into a multi-hyphenate AI builder, and the practical agentic workflows she uses every day to reclaim her time .*⏳ Timestamps:* *0:00* - Introduction: Leading AI at Google Maps and mentoring 100+ AI founders *0:27* - Product thinking for 1 user vs. 1 billion users *1:10* - How LLMs changed data analysis and product architecture *2:23* - Human alignment is the new bottleneck in tech *3:30* - The shift from 5 decisions a week to 20 decisions a day *3:37* - Why Subject Matter Experts (SMEs) are taking over leadership decisions *4:00* - Automation vs. Human-in-the-Loop: An AI reporting case study *5:55* - Why AI is a database, not a thinker *6:23* - Will TPM, PM, and Engineering roles merge into one? *7:45* - How AI makes product leaders proactive instead of reactive *9:33* - How to stay relevant and "add value + 1" in an AI-driven market *11:00* - Transitioning careers: From Admin or UX Designer to PM using AI *12:00* - Anjali's AI stack: Gemini, NotebookLM, Claude, and Anti-Gravity *13:50* - Predictions: Agents will own everything on Google Maps in 2 years *15:00* - How Anjali uses daily AI agents to filter emails, pings, and tasks *16:23* - Final thoughts: "No one can take your job but you" *Anjali:* [  / anjalikakkadp  ](  / anjalikakkadp  )*Leena:* [  / leenaagarwal  ](  / leenaagarwal  )*🔗 Connections:*

  7. 7 Jul

    60% of Web Traffic is AI: Why You Must Stop Building for Humans

    **The internet has fundamentally changed.** Over 60% of web traffic is now driven by AI agents, not humans. If you are still building landing pages and products exclusively for people, you are missing the biggest shift since the dot-com boom.In this episode of *Agentic AI Execution*, Mariane—Head of DevRel at You.com and founder of the Founders Bay community—breaks down why we are entering an AI renaissance rather than a bubble. We discuss why developers must evolve into "full-stack AI engineers" who manage 10-20 agents simultaneously to stay competitive. We also dive deep into the reality of building a massive community from scratch, the failure patterns of early-stage founders building in isolation, and the very real balancing act of scaling a career while managing mom guilt.**⏳ Timestamps:**0:00 - Over 60% of web traffic is now AI agents0:37 - Mariane's journey: Engineering leadership to You.com DevRel2:21 - The #1 reason AI founders fail: Building in isolation4:47 - Why you need to build websites for agents, not humans6:13 - AI Bubble vs. The AI Renaissance7:45 - Vibe Coding: Managing 10-20 agents at once9:11 - Handling bugs and CI/CD with AI-generated code11:21 - The rise of the Forward Deployed & Full Stack AI Engineer14:50 - Growing Founders Bay and building community from scratch18:30 - How to build an audience from zero today22:15 - The future of DevRel in an Agentic AI world23:30 - Managing burnout, parenting, and the reality of "mom guilt"26:06 - Closing thoughts: Keep building and showing up🔗 Connections:* Mariane: https://www.linkedin.com/in/marianebekker/ * Leena: https://www.linkedin.com/in/leenaagarwal/* Founders Bay Community: https://www.linkedin.com/company/foundersbaynetwork

  8. 30 Jun

    Why AI Agents Still Make Wrong Calls — Even With Perfect Memory | Adelina, Founder of Rory Plans

    What if the hardest problem in AI agents isn't memory — it's judgment?In this episode of Agentic AI Execution, I sit down with Adelina — former colleague at Cruise, later at Vercel, and now founder of Rory Plans, building AI agents for organizational decision-making.We get into why agents with perfect memory still make the wrong recommendation, why bad decisions compound faster than you'd think, and what she wishes she'd known before quitting her job to start her company.In this conversation:→ Why memory tells an agent what happened — but not what matters→ What Cruise taught her about decisions under uncertainty, and how it applies to AI agents today→ Why Rory Brain exists to capture judgment, not just information→ Running "Lunch and Hack" events and what builders are getting wrong with AI agent memory→ Why she believes organizational intelligence — not bigger models — will define the winners in AI→ Her advice to founders: spend 80% of your time with customers, 20% building🔗 Connections:* Adelina: https://www.linkedin.com/in/adelinamartiniuc/* Leena: https://www.linkedin.com/in/leenaagarwal/* Rory Plans: https://roryplans.aiTimestamps:00:00 Intro00:26 Adelina's journey: Cruise → Vercel → founding Rory Plans02:30 Why judgment is harder than memory for AI agents05:00 What Cruise taught her about decisions at scale07:30 Building Rory Brain to capture organizational judgment10:30 Lunch and Hack: what builders get wrong about agent memory13:30 Why organizational intelligence will define AI's winners17:00 Advice for founders: customers over building20:05 Closing takeaways + where to find Adelina#AgenticAI , #AIAgents , #TPM , #AI

See All (21)

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
    2026
  • Episodes
    21
  • Rating
    Clean
  • Copyright
    © Leena Agarwal
  • Show Website
    Agentic AI Execution