The Spark & The Forge: Patterns That Actually Work

Subrata Kar
  • 0.0 (0)
  • Business
  • Weekly

Extracting scaling patterns from 200+ conversations with leaders building AI, healthcare, and enterprise systems. This isn't theory—it's what's working in production. Each episode breaks down decisions, frameworks, and approaches practitioners used to scale successfully (and what failed). Topics: AI governance, enterprise scaling, product-market fit, GTM strategy, leadership hiring, deeptech innovation. Host: Subrata Kar, mentor at T-Hub & NASSCOM DeepTech Club, 40 years in enterprise systems. For founders and CTOs who need specifics, not buzzwords.

  1. Oct 7

    What Does “Doctor in Your Pocket” Really Mean for Healthcare?

    What stayed with me after recording Episode 86 with Dr. Anand Iyer? He described the idea of having your “doctor in your pocket” — something that sees what is happening and guides you through it, moment by moment. But the more I thought about our conversation, the more complicated that idea became. Diabetes is well codified. If glucose is X, the time of day is Y, and the last reading is Z — do this. You can take that clinical logic and build a rules-based system around it. But what happens when we move to conditions we don't understand or can't yet codify? And then there is another question. We don't come with one condition at a time. We can have diabetes, hypertension and heart failure all at once. We don't experience those as three separate applications. We experience one body. So can digital health understand how those conditions interact — and personalize the support around the person? Anand also made an important point about N=1. Even when you start with population-level guidelines, the person in front of you is still an individual. The question is how you personalize within those guidelines — what actually works for that particular person. We also got into GLP-1s. I asked Anand directly whether a drug this effective could make a coaching application unnecessary. His answer was more nuanced. His idea of the app as a navigation aid through the GLP-1 journey made sense to me: Helping someone understand what they're taking.Helping them deal with the side effects.Reminding them what they need to do.Helping them stay on track. The molecule is the molecule. It doesn't change. But that layer around it can help the person actually get the benefit from the treatment. And that's what I kept thinking about after the recording. Digital health isn't one breakthrough away from being solved. It's not just the smarter model, the newer drug or the bigger platform. It's more nuanced than that — disease by disease, condition by condition, person by person. Maybe the real frontier isn't just making AI smarter. Maybe it's making the support more relevant to the person actually living with the condition. That's the question I'm still thinking about. This is my reflection on Episode 86 with Dr. Anand Iyer, Chief AI Officer at Welldoc. ▶️ Watch E86 — The full conversation with Anand Iyer:https://youtu.be/y_2nThq2tNk Why diabetes was a natural starting point for AI-driven health coachingWhat happens when a disease isn't yet well codifiedWhy N=1 matters in personalized healthcareThe challenge of managing multiple conditions togetherGLP-1s and the role of digital coachingWhy the future of digital health may be about relevance, not just intelligence Subrata KarHost — The Spark & The Forge I study patterns from builders who scale — enterprise systems, AI platforms, and startups — and extract actionable insights leaders can apply immediately. LinkedIn:https://www.linkedin.com/in/subroto Newsletter:https://substack.com/@subratakar What do you think? Can AI become a meaningful layer of support around healthcare — especially when people don't experience their conditions one at a time? What you'll hearAbout Subrata KarJoin the conversation#AIHealthcare #DigitalHealth #AnandIyer DISCLAIMER This video is for informational and educational purposes only. It does not constitute medical advice and should not be used as a substitute for advice from a qualified healthcare professional. Discussions of diabetes, medications, healthcare, artificial intelligence and digital health are for informational purposes. Always consult qualified healthcare professionals before making healthcare decisions. Visual infographics and data visualizations may have been generated or assisted by AI tools for illustrative purposes. FAIR USE NOTICE Charts, news clips and referenced materials are used for purposes of commentary, criticism, analysis and education under Section 107 of the Copyright Act #HealthAI #GLP1 #TheSparkAndTheForge

  2. Oct 5

    AI Health Coaching: Your Doctor in Your Pocket?

    What happens when AI becomes your health coach — learning from your data and helping you make decisions in real time? Dr. Anand Iyer describes the idea as having your “doctor in the pocket.” But how much should we trust it? In E86 of The Spark & The Forge, I sit down with Dr. Anand Iyer, Chief AI Officer at Welldoc, to explore what happens when AI moves from simply showing us information to coaching us in real time. The conversation starts with diabetes and the data already available to patients — medications, education, diet, activity, labs and symptoms — and then goes much further. Anand explains how continuous data and rules-based systems can help patients see cause and effect in their own lives. One patient described the experience this way: “I’m allowing myself to hack my own body.” One slice of pizza versus two. Medication on time versus late. What happened this week compared with last week. The idea is not simply that AI tells you what to do. It can help people understand their own patterns and behavior — what Anand describes as N equals one. We then get into the harder questions. How different is rules-based AI from generative AI? How close should AI get to medication decisions? Can patients become too dependent on an AI health coach? What happens when clinical trial conditions disappear and the system has to work in the real world? Why does codification matter so much in healthcare? And if AI gets closer to decisions that affect a person's health, what does trust actually mean? Clinical outcomes. Safety and regulation. Cybersecurity. Usability. And finally, accountability. If an AI recommendation contributes to a bad outcome, who is actually on the hook? This is a conversation about AI in healthcare — but the bigger question goes beyond healthcare: What happens when AI moves from giving us information to influencing decisions that actually matter? Chief AI Officer — Welldoc Dr. Anand Iyer is Chief AI Officer at Welldoc, working at the intersection of artificial intelligence, digital health and patient care. Diagnosed with Type 2 diabetes in 2002, Anand brings a particularly personal perspective to the conversation — he has lived with the condition while helping build technology designed to support people managing it. LinkedIn:https://www.linkedin.com/in/anand-iyer-b294b64/ Host — The Spark & The Forge I explore the ideas, decisions and patterns that matter to founders, technology leaders and builders working with enterprise technology, AI and startups. LinkedIn:https://www.linkedin.com/in/subroto/ Newsletter:https://substack.com/@subratakar If an AI system is coaching you continuously using your own health data, where should the line be between: Prediction → Recommendation → Coaching → Action? Where would you draw it? Spotify:https://open.spotify.com/show/5Ck1ZktTrloA579kVTzxg0 #AIHealthCoaching #AIHealthcare #AnandIyer #Welldoc #DiabetesManagement #TheSparkAndTheForge DISCLAIMER This video is for informational and educational purposes only. It does not constitute medical advice and should not be used as a substitute for advice from a qualified healthcare professional. Discussions of diabetes, medications, healthcare, artificial intelligence and digital health are for informational purposes. Always consult qualified healthcare professionals before making healthcare decisions. Visual infographics and data visualizations may have been generated or assisted by AI tools for illustrative purposes. Charts, news clips and referenced materials are used for purposes of commentary, criticism, analysis and education under Section 107 of the Copyright Act. FAIR USE NOTICE Charts, news clips and referenced materials are used for purposes of commentary, criticism, analysis and education under Section 107 of the Copyright Act.

  3. Oct 3

    When AI Becomes the Enterprise: Who's in Control? | E89

    What happens when AI stops being a tool you use and starts becoming part of how the enterprise operates? Enterprise software is already a landscape — ERP, CRM, HR, procurement, finance, service and industry-specific systems. AI agents can increasingly operate across that landscape, interpret information, make decisions and take actions. That changes the question. It's no longer just whether AI can do the work. It's whether the enterprise can control the intelligence doing it — switch it, understand it, govern it, reconstruct what happened, and keep operating when it isn't available. In this episode, I explore what happens when AI becomes an operating dependency inside the enterprise. WHAT I EXPLORE • What changes when AI agents start acting across enterprise systems• Why technical portability is not the same as operational portability• What happens when different agents or models make different judgments• Where intelligence lives when it is distributed across applications and platforms• Who is actually in control when intelligence is embedded inside the enterprise stack• The five things enterprises need to understand: Data, Decisions, Intelligence, Provenance and Resilience• Why probabilistic intelligence can operate inside hard boundaries• Why a perfectly enforced rule can still be the wrong rule• Why systems of record still matter as systems of agents emerge• The seven-day test: what happens if the AI your business depends on simply disappears?• When AI stops being a capability and becomes an operating dependency THE CORE QUESTION The enterprise landscape is becoming more intelligent piece by piece. But the harder question may not be how intelligent the enterprise becomes. It's how much of the enterprise we are willing to make dependent on intelligence that we don't control. ABOUT THE HOST Subrata KarHost — The Spark & The Forge I study patterns from builders who scale — enterprise systems, AI platforms, and startups — and extract actionable insights leaders can apply immediately. LinkedIn: https://www.linkedin.com/in/subrotoNewsletter: https://substack.com/@subratakar JOIN THE CONVERSATION If AI becomes part of how your enterprise operates, what would you need to know before you were willing to depend on it? #EnterpriseAI #AIGovernance #EnterpriseArchitecture #AIAgents #AgenticAI #TheSparkAndTheForge #SubrataKar DISCLAIMER: This episode is for informational and educational purposes only. It does not constitute professional financial, investment, or legal advice. Visual infographics and data visualizations may have been generated or assisted by AI tools for illustrative purposes. Always conduct your own research and consult with qualified professionals before making business, legal, or financial decisions. FAIR USE NOTICE: Charts, news clips, and referenced materials are used for purposes of commentary, analysis, criticism, and education under Section 107 of the Copyright Act (Fair Use).

  4. Oct 2

    The Future of ERP: What Actually Works With AI | E88

    AI is changing how enterprise software gets built. But what does that mean for ERP? The ERP problem has changed. The old problem was customization — and the technical debt that came with it. SaaS standardized the core, but enterprises increasingly became landscapes of systems: SAP or Oracle for finance, Salesforce for CRM, Workday for HR, procurement platforms and industry-specific applications. Most of those systems are good at what they do. The harder problem is what happens between them. A business process doesn't care where the software boundary is. Today, very often, the human being is the integration layer. I think AI changes that. Instead of one giant AI trying to become the whole ERP, we could see specialized agents assembled around existing systems of record — connecting applications, building new workflows and increasingly interacting with agents outside the enterprise. Your procurement agent talking to a supplier's agent. Your AP agent talking to another company's AR agent. The potential shift is bigger than automation. It could compress the friction between companies from days to minutes. But does that mean the core systems disappear? I don't think so. The system of record still matters. The ledger matters. Compliance matters. And when something goes wrong, someone has to own the outcome. AI can dramatically reduce the cost of understanding old systems, connecting them and assembling new workflows. It doesn't automatically solve the organizational problem underneath them. The process may still need to change. Someone still needs to decide what the right process is. And someone still has to be accountable when AI gets it wrong. That is where I think the Trust Premium comes in. The value doesn't disappear when execution becomes cheap. It moves to whoever can govern the outcome and stand behind it. WHAT I EXPLORE • Why the old ERP customization problem has changed• Why the enterprise application landscape matters more than “the ERP”• Why humans have become the integration layer• How AI changes application assembly• Systems of record vs. systems of agents• Why vertical domain expertise could become the moat• What agent-to-agent interaction could mean for enterprise processes• Why AI doesn't automatically solve process and accountability REFERENCED CONVERSATIONS Malcolm Hawker — Former Gartner Analyst | Master Data Management & Data StrategyLinkedIn: https://www.linkedin.com/in/malhawker Luis Lamb — AI, Tech & Innovation Leader | Researcher | Educator | Neurosymbolic AILinkedIn: https://www.linkedin.com/in/luis-lamb-131394 ABOUT THE HOST Subrata KarHost — The Spark & The Forge I study patterns from builders who scale — enterprise systems, AI platforms, and startups — and extract actionable insights leaders can apply immediately. LinkedIn: https://www.linkedin.com/in/subrotoNewsletter: https://substack.com/@subratakar JOIN THE CONVERSATION If AI becomes the layer connecting enterprise systems, what does ERP become? #ERP #EnterpriseAI #AgenticAI #AI #ERPTransformation #TheSparkAndTheForge #SubrataKar DISCLAIMER: This episode is for informational and educational purposes only. It does not constitute professional financial, investment, or legal advice. Visual infographics and data visualizations may have been generated or assisted by AI tools for illustrative purposes. Always conduct your own research and consult qualified professionals before making business, legal, or financial decisions. FAIR USE NOTICE: Charts, news clips, and referenced materials are used for purposes of commentary, analysis, criticism, and education under Section 107 of the Copyright Act (Fair Use).

  5. Sep 1

    Does AI Really Understand? | Luis Lamb | E87

    Does AI really understand what it writes — or can it produce an answer that looks like understanding? Prof. Luis Lamb and I start with a deceptively simple question: how do we know whether AI has given us knowledge — or just a very good guess? From there, we explore what happens when AI moves from answering questions to taking actions. We talk about reasoning, neurosymbolic AI, autonomous agents, compounding errors, guardrails and verification — and a harder question: even if we can verify that an AI system followed a rule, how do we know the rule itself is right? And underneath all of it is a bigger question: As AI becomes more capable and more autonomous, how do we make sure it remains beneficial to humanity? Prof. Luis Lamb is a researcher, educator and pioneer in neurosymbolic AI. His career spans AI research, academia, industry and government, including leadership roles at Boeing and technology and innovation work in Brazil. LinkedIn: https://www.linkedin.com/in/luis-lamb-131394/ The Spark & The Forge Hosted by Subrata Kar — conversations with builders, researchers and technology leaders about the ideas shaping how we build and deploy AI. LinkedIn: https://www.linkedin.com/in/subroto Newsletter: https://substack.com/@subratakar DISCLAIMER ⚠️ DISCLAIMER: This video is for informational and educational purposes only. It does not constitute professional financial, investment, or legal advice. Visual infographics and data visualizations in this video may have been generated or assisted by AI tools for illustrative purposes. Always conduct your own research and consult with qualified professionals before making business, legal, or financial decisions. 📜 FAIR USE NOTICE: Charts, news clips, and referenced materials in this video are used for purposes of commentary, analysis, criticism, and education under Section 107 of the Copyright Act (Fair Use).

  6. Jul 18

    Who Decides? AI, Accountability & the Hidden Choices Behind Every AI System | Prof. Renée Cummings

    Who decides when an algorithm gets it wrong? That's the question at the heart of this conversation with Professor Renée Cummings—a criminologist who spent two decades in criminal justice before becoming one of the world's leading voices in AI governance. Every day, AI systems influence decisions that affect people's lives: who gets hired, who gets approved for a loan, who receives medical care, and who gets flagged as a potential risk. We often assume these systems are objective because they're built on data. But data carries history—and history carries bias. Renée explains what happens when that history becomes training data, why nobody walks away clean when an algorithm causes harm, and why she believes some AI systems should never be fixed—they should be deleted. This isn't a theoretical discussion. It's about the decisions increasingly being made about our lives by systems we rarely see and often cannot question. • What "data trauma" means and how historical bias becomes part of AI systems• Why algorithms can amplify past mistakes at unprecedented speed and scale• The real accountability chain when AI causes harm—and why responsibility doesn't stop with developers• When a broken AI system should be fixed versus removed entirely• Why Renée says "Code is the new cop"• How AI is changing hiring, lending, healthcare, criminal justice, and public trust• The emerging risks of algorithmic grooming, AI companions, and child safety• Why every AI system is ultimately a reflection of human values• What founders, product leaders, and CTOs should consider before deploying AI into high-impact decisions• How governance must evolve as AI moves from assisting humans to acting autonomously Professor Renée Cummings is Professor of Practice at the University of Virginia School of Data Science, Nonresident Senior Fellow at the Brookings Institution, former member of the World Economic Forum's Data Equity Council, member of the Global AI Governance Alliance, Founder of A.W.A.R.E. (Algorithmic Wellbeing & Responsible Engagement), and is recognized among the world's Top 100 Women in AI Ethics. She has advised policymakers, including the U.S. Senate, bringing a criminologist's perspective to AI governance, ethics, and public policy. Subrata Kar is the host of The Spark & The Forge, where he explores how founders, researchers, policymakers, and enterprise leaders are building the future of AI through thoughtful, long-form conversations that go beyond headlines and hype. Listener Advisory: This episode includes discussion of child safety, emotional harm, and youth suicide in the context of AI companion technologies. Listener discretion is advised.

  7. Jun 16

    Your AI Bill Just Quadrupled — Here's Why

    Your AI contract was written for a per-seat world. The meter is running in a token world. Microsoft reversed course on Claude Code. Uber burned through its entire 2026 AI coding budget in four months. Both for the same reason — token-based billing scales faster than anyone budgeted for. Four pricing models are now running simultaneously in enterprise AI. In almost every one, the buyer carries more risk than the headline price suggests. This episode breaks down all four — and the one question you must ask before every AI renewal conversation. Host:Subrata Kar — enterprise technology strategist and host of The Spark & The Forge. I study patterns from builders who scale and extract insights enterprise leaders can apply immediately. LinkedIn: https://www.linkedin.com/in/subrotoNewsletter: https://substack.com/@subratakarYouTube: https://youtube.com/@thesparkandtheforge Disclaimer:This episode is for informational and educational purposes only. It does not constitute professional financial, investment, or legal advice. Always conduct your own research and consult with qualified professionals before making business or financial decisions.

  8. May 29

    AI Is Changing Engineering Jobs Fast | Brian Samson

    AI is changing software engineering faster than most teams are prepared for. AI coding tools are generating code faster than engineers can review it. Junior hiring is slowing down. And the value of engineering work is shifting from pure execution toward validation, judgment, ownership, and context. In Episode 75 of The Spark & The Forge, Subrata Kar speaks with Brian Samson — founder of Plugg Technologies and recruiter placing engineering talent across Latin America and US tech companies. This conversation explores: why CTOs are still hiring despite AI coding tools the hidden “rework tax” behind AI-generated codewhy demand is shifting toward senior engineersdeepfakes and bots in technical interviewsIndia’s engineering transitionthe rise of the “full stack unit”TIMESTAMPS This is not a doom conversation about AI replacing engineers. It is a conversation about what becomes more valuable as AI improves. 00:00 — “I think that company is not gonna make it”00:16 — The shift nobody is naming clearly yet01:08 — Why CTOs are still hiring despite AI coding tools01:54 — The hidden rework tax of AI-generated code02:44 — Why demand is shifting toward senior engineers03:37 — Bots and ghost candidates in interviews06:33 — What this means for India’s engineering pipeline08:38 — Why YC startups are hiring engineers in Latin America12:19 — “Become the full stack unit”13:27 — The window is still open. But not indefinitely. Brian Samson https://www.linkedin.com/in/briansamson/ Founder — Plugg TechnologiesHost — The Nearshore Cafe Podcast Subrata Kar https://www.linkedin.com/in/subroto/Host — The Spark & The Forge GUESTHOST

See All (22)

About

Extracting scaling patterns from 200+ conversations with leaders building AI, healthcare, and enterprise systems. This isn't theory—it's what's working in production. Each episode breaks down decisions, frameworks, and approaches practitioners used to scale successfully (and what failed). Topics: AI governance, enterprise scaling, product-market fit, GTM strategy, leadership hiring, deeptech innovation. Host: Subrata Kar, mentor at T-Hub & NASSCOM DeepTech Club, 40 years in enterprise systems. For founders and CTOs who need specifics, not buzzwords.

Information