Welcome to episode 29 of the Agentic AI Execution podcast! In this episode, Leena sits down with Akshay, a Google Developer Expert (GDE) for Android, Developer Advocate, and founder of the Android Engineers Community in India.

Akshay shares his journey of building with Agentic AI, starting with a unique digital Ganesh Pandal powered by voice AI and GPT Live 1. The conversation dives deep into the realities of developer relations (DevRel), how to effectively use Gemini inside Android Studio, and the critical differences between a flashy AI demo and a resilient model that survives real-world production edge cases.

**Key Topics Covered:**

* **00:00** – Introduction and exploring the real gap between a flashy AI demo and one that survives production.

* **01:01** – Demo: Building a digital Ganesh Pandal with voice AI so anyone can join the festival virtually.

* **03:15** – The path to becoming a GDE and the importance of advocating for developers.

* **08:00** – Leveraging Gemini AI in Android Studio for faster coding and documentation.

* **10:30** – Protecting developer trust and providing quick-start AI templates via Agora

.* **12:50** – Real-world lessons: What to do when your live tech demo breaks.

* **22:14** – Navigating latency, background noise, and creating natural human-like voice agents

.

* **24:45** – Demos vs. Production: Why handling edge cases like low lighting and choppy video matters.

* **27:10** – The massive untapped opportunity for Voice AI in children's toys across India.

🔗 RESOURCES & LINKS:

* Connect with Akshay on LinkedIn: https://www.linkedin.com/in/anandwana001/

* Connect with Leena on LinkedIn: https://www.linkedin.com/in/leenaagarwal/

In Episode 28 of Agentic AI Execution, host Leena Agarwal sits down with James (Senior PM, former Staff TPM, and author of The TPM Craft newsletter) to unpack how agentic workflows are reshaping technical leadership roles across startups, mid-sized companies, and large enterprises.James breaks down why human relationships matter more than ever, where trust still breaks down in agentic code generation (security, compliance, quality), and the exact skills TPMs must double down on to stay indispensable. Plus, he shares real-world projects he built using AI—including a short-form content blocker and a virtual bank app for kids!

⏱️ CHAPTERS / TIMESTAMPS:

• 0:00 – Intro: Setting the Stage for the Agentic AI Era

• 0:45 – Roles, Trust, & The Secret Sauce for Using AI Well

• 21:13 – Scarce Skills TPMs Need as Execution Gets Cheaper

• 35:41 – Closing Thoughts

🔗 RESOURCES & LINKS:

• Subscribe to James’s Newsletter (The TPM Craft): https://thetpmcraft.substack.com/

• Connect with James on LinkedIn: / jamzdayhuff

• Connect with Leena on LinkedIn: / leenaagarwal

In Episode 27 of the Agentic AI Execution podcast, host Leena sits down with Technical Program Manager and coach Liudmila to uncover how AI is doubling expected workloads in the tech industry. They dive into the reality of "bot-sitting" (spending hours fixing AI-generated mistakes), practical neuroscience-backed tools to stay grounded under pressure, and how to build sustainable execution systems without sacrificing your mental health.

Key Topics Covered:

* 00:00 – The hidden reality of AI expectations and tech worker fears

* 01:36 – Liudmila’s career journey from test engineer to TPM

* 16:45 – Real-time grounding exercises to manage pressure and anxiety

* 24:30 – Is AI boosting productivity or creating 6.4 extra hours of "bot-sitting"?

* 31:10 – Managing up and reframing career value in the era of AI agents

* 35:38– Final advice on prioritizing life outside of work and avoiding burnout Links:

* Connect with Liudmila: https://www.linkedin.com/in/liudmila-gishman/ * Connect with Leena: https://www.linkedin.com/in/leenaagarwal/

In this episode AWS Community Peter break down a holistic framework for AI adoption. Discover why top-down AI mandates fail, how engineering and product teams must adapt their specifications, and why standard ROI metrics are measuring the wrong things.If your team is trying to figure out how to successfully roll out agentic AI workflows without sacrificing software reliability, this conversation is a must-watch.

**What You'll Learn in This Episode:***

Why AI adoption is a holistic culture and process transformation, not just a tooling problem

* Practical guardrails and linting rules to prevent AI-generated "code slop"

* The real truth about measuring AI ROI and developer productivity metrics

* The single biggest change engineering leaders should make on Monday morning

**Episode Timestamps:***

**00:00** – Introduction: Meet Peter and the Agentic AI Execution Podcast*

**01:10** – The People, Process, and Technology Framework for AI Adoption*

**05:30** – Best Practices for Deploying Quality AI Agents*

**08:40** – How to Prevent AI Slop and Handle Tech Debt*

**13:00** – Measuring True AI Return on Investment (ROI)*

**16:00** – Production-Ready Agentic AI: Startups vs. Enterprise Realities*

**17:50** – One Thing Engineering Leaders Must Change This Monday Morning*

**18:52** – Key Takeaways and Final Thoughts

🔗 Links:

Connect with Peter: https://www.linkedin.com/in/petersankauskas/

Connect with Leena: https://www.linkedin.com/in/leenaagarwal/

If your team is still running Story Points and Fibonacci on AI projects, today's guest says that math is already broken.In this episode, from actionable frameworks to the hard truths about AI ROI, Shivam argues that before you implement the technology, you absolutely must fix your broken processes first. We also dive into why we need "human in the *line*" instead of just "human in the *loop*," and exactly what a 30-60-90 day AI integration plan looks like for a mid-sized company.If you are navigating the chaos of AI, this is a must-listen episode!---### ⏱️ **Chapter Timestamps*** **3:31** - The Biggest Bottlenecks: Why agents fail to reach production* **5:45** - Surviving the Chaos: How the TPM role must evolve* **10:03** - The AI Truth: Balancing cost savings, productivity, and employee experience* **12:24** - The #1 reason Agentic AI pilots die (and why you need a charter)* **14:28** - "Human in the Line": Why you need to fix your processes before deploying AI* **19:22** - The Ultimate 30-60-90 Day AI Action Plan for mid-sized companies* **24:50** - Why traditional capacity tracking and methodologies belong out the door* **25:58** - Rethinking AI Security & handing over the keys to the kingdom* **28:20** - Essential skills PMs need for the Agentic AI era* **31:15** - 18 Months Out: Preparing for your new "L3" AI Co-worker* **33:00** - Finding our "uniquely human" purpose in the age of AI* **35:04** - Outro & Final Thoughts🔗 Links:* Connect with Shivam: https://www.linkedin.com/in/shivamkhullar/* Connect with Leena: https://www.linkedin.com/in/leenaagarwal/#AgenticAI #AgenticAIExecution

She's evaluated hundreds of AI pitch decks — and she can spot a real agentic AI company from a wrapper in two sentences.

In Episode 24, I sit down with Sonia, Managing Partner at Zentara Ventures and founder of Dekode Health, to unpack what actually separates AI companies with real ROI from the ones wrapping a chatbot around a spreadsheet and calling it a platform.Sonia isn't theorizing from the sidelines. She's an engineer by training who moved through product and marketing before landing on the venture side — and she's personally built and launched her own AI healthcare product, now in private pilot. Four exits, one IPO, and more than half a billion dollars raised as a C-level executive — she's built a due-diligence lens most investors don't have.

⏱️ Chapters:0:00 Intro

0:32 Sonia's Path: Engineer to Product to Marketing to VC

2:37 What "AI-Powered" Really Buys You in a Pitch5:09 Real AI vs. Wrapper: The VC Due Diligence Test

6:12 Why She Went Deep Before Going Wide

8:23 The 91% Problem: Marketers Faking Agentic Work10:45 The Two-Sentence Test for Fake Agentic A

I13:06 Legal Exposure: Where Founders Overstate AI14:04 What Should Never Get Handed to an Agent

15:36 Agentic AI in Marketing: Why Humans Still Win

17:34 The Line Healthcare AI Should Never Cross

19:42 The ROI Paradox: 84% Positive vs. "It's All Hype"

22:10 AI Layoffs: Real Capability or Convenient Cover?

25:35 The Solopreneur Era (and Why You Still Need Humans)

27:04 Redefining Execution in 2026

29:42 What Changes by 2027

32:15 Where to Find Sonia + Dekode Health

"Founders who treat AI as a marketing layer instead of an infrastructure layer — wrapping a ChatGPT or Claude API around a spreadsheet and calling it a platform — that's not a company. It's a feature at best." — Sonia

🔗 Links:* Connect with Sonia: https://www.linkedin.com/in/soniahunt/

* Connect with Leena: https://www.linkedin.com/in/leenaagarwal/

Subscribe for new episodes of Agentic AI Execution, where we go past the pitch deck to how AI actually gets built, funded, and used.

#AgenticAI #AgenticAIExecution #VentureCapital #AIStartups #HealthcareAI

Are you throwing money into a black hole with paid ads? Deepti, founder of Uplaud and former product lead for Meta's monetization engine, explains why the future of growth isn't more ad spend—it's customer trust.In this episode of Agentic AI Execution with Leena Agarwal, we dive into why customer referrals convert 5x better than paid media, how agentic AI is transforming the way businesses acquire users, and why the tech stack of the future relies on authentic voices, not just algorithms.

⏱️ Chapters:

0:00 - Intro: Customer Trust is the New Gold

1:13 - Deepti's Journey: From Meta to Startup Founder

4:50 - The Problem with Modern Ad Platforms

8:15 - Why 5-6% is Still a "Good" Conversion Rate

11:30 - How Uplaud Amplifies Authentic Customer Voices

16:45 - Uplaud's AI Stack: Claude, OpenAI, and Gemini

18:20 - Career Advice for PMs & Engineers in the AI Era

🔗 Links:

• Check out Uplaud: https://www.uplaud.ai/
• Connect with Deepthi: / deepthirao05
• Connect with Leena: / leenaagarwal

Aadil has spent 16 years as a TPM across Apple, Google, Nike, and Humane — and he's convinced the biggest risk in the AI era isn't using these tools too little. It's reaching for them before asking whether you even need to.

🕒 Timestamps:

00:00 - Intro & The Airtable Epiphany

00:53 - Why Good Judgment Beats Good Prompts

03:15 - Catching Dilemmas Early & Avoiding Agent Maintenance Overhead

05:10 - What to Delegate vs. What to Keep Control On

08:30 - TPMs Are Synthesizers, Not Producers

11:00 - How TPMs Stand Out in the AI Era

15:30 - What Hiring Managers Look For in Senior TPMs

19:00 - Startup Thinking vs. Big Tech Process Bloat

23:15 - Driving Velocity with Repeatable Systems

25:40 - Feedback Loops & Leadership Lessons from Apple 27:36 - When AI Agents Fail & Avoiding "Judgment Debt" 30:15 - Final Career Advice: Never Stop Learning

32:50 - Key Takeaways & Wrap-Up

🔗 Resources & Links:*

* Aadil writes a newsletter on TPM leadership — https://artoftpm.substack.com/

* Connect with Aadil: https://www.linkedin.com/in/aadilmaan/

* Connect with Leena: https://www.linkedin.com/in/leenaagarwal/