The Integration Layer

Shubhendu Tripathi

Everyone is talking about AI. Almost nobody is telling the truth about it. This is the show where builders, operators, and the people quietly reshaping enterprises sit down and say what they actually think. No hype cycles. No buzzword bingo. Just honest conversations about what AI is really doing to our companies, our jobs, and the way we make decisions. I'm Shubhendu. I build AI systems for a living, and I've seen what works, what breaks, and what nobody wants to admit in the boardroom. Every week I bring on the people behind the systems, the strategies, and the bets that are defining this decade, and I ask them the questions everyone is thinking but rarely says out loud. If you want the polished keynote version, there are a thousand of those. If you want the real version, pull up a chair.

Episodes

  1. Oct 1

    Pharma Spent 30 Years Building Channels. AI Reads .gov Instead

    The conversation explores the impact of AI on various industries, particularly focusing on the experiences of Leandra Wells, who has utilized AI tools to build her own CRM and enhance her business operations. It discusses the challenges and opportunities presented by AI, especially in regulated environments like pharmaceuticals, and emphasizes the importance of adapting to technological advancements while maintaining accuracy and credibility in information dissemination. The conversation explores the evolving role of AI in healthcare, the regulatory landscape, and the importance of change management in adopting new technologies. Leandra Wells and Shubhendu discuss the challenges and opportunities presented by AI, emphasizing the need for governance and responsible use. Takeaways AI democratizes technology accessImportance of problem-solving skills AI adoption requires careful governanceChange management is critical for success Chapters 00:00 The AI Debate: Truths and Myths03:11 Leandra's Journey with AI Tools05:55 Building a CRM: A Non-Techie's Perspective08:56 Reflections on Career and AI's Impact11:47 Gender Disparities in AI Adoption15:08 The Future Workforce: Adapting to AI17:58 The Shift from Google to AI for Information20:54 Navigating AI in Regulated Industries23:48 The Challenge of Credible Information in AI27:00 Trust in AI vs. Traditional Sources30:03 Getting Started with AI in Business32:56 Pharma's Strategy in the Age of AI37:40 Navigating AI in Healthcare39:54 Regulatory Challenges and AI42:52 The Role of IT in AI Implementation46:00 Debunking AI Failure Myths49:53 Change Management in AI Adoption53:49 AI as a Tool for Efficiency56:46 Environmental Impact of AI01:00:52 Future of Work with AI01:04:58 Governance and Responsibility in AI

  2. Sep 24

    Feeling Is Not a KPI | AI You Can Actually Prove

    Everyone is talking about AI. Almost nobody is telling the truth about it. Lee Crossley builds tools whose main job is to demonstrate that the thing everyone is buying does not do what they were told it does. He spent years in hardware, ran technical quality for roughly 2,500 engineers across 13 offices, and today writes the audio drivers himself for a product he is building with one other person. His argument in this episode is that AI models have plateaued into a commodity, and that every decision still worth arguing about now sits in the layers above them. What You Will Learn Why AI text watermarking collapses the moment someone paraphrases the output, and what a real provenance trail would have to record insteadThe Monday morning question every CIO should put to an AI vendor, which exposes in one answer whether a rollout is real or theatreHow to measure an AI productivity claim so that it survives a conversation with your CFOWhy only 5 percent of enterprise AI pilots produce measurable business impact, and why the research points at learning rather than model quality or regulationWhat a memory layer actually is, why providers over-send your context, and what that costs you in privacy and tokensHow local speech to text closed the accuracy gap with the cloud, and why that changes what regulated teams are allowed to buildWhat Lee looks for when hiring now that CVs are written by AI and screened by AIAbout the Guest Lee Crossley is a consultant CTO, a director at San Digital and co-founder of Overshow, a privacy-first work memory for macOS that keeps screen context, meeting audio and notes on the user's own machine. He was previously Chief Engineer at BJSS and CTO of a hardware startup that took him to Taiwan around 20 times in three years. He is lead author of Production-Grade AI: From Boardroom to Build and Beyond, and he teaches AI-assisted engineering and AI-assisted knowledge work through San Digital.

  3. Aug 20

    AI Is an Unforgiving Magnifying Glass

    The conversation with Dr. Sebastian Wernicke delves into the challenges of being data-driven versus data-inspired, the role of culture in decision-making, and the limitations of dashboards in reflecting the true state of a business. Dr. Wernicke emphasizes the need for a shift from data-driven to data-inspired decision-making and the cultural barriers that hinder this shift. The conversation with Dr. Sebastian Wernicke delves into the impact of AI on data culture and the role of data scientists in the age of AI. It emphasizes the importance of purpose-driven AI and the limitations of technology in fixing broken cultures. Dr. Wernicke also discusses the misinterpretation of data and the significance of human skills in data science. Takeaways Data-driven decision-making optimizes existing processes, while data-inspired decision-making seeks to transform and innovate.Culture plays a significant role in decision-making, and the use of data is often influenced by cultural factors within an organization. Purpose-driven AI overcomes broken data cultureData scientists bring analytical and business skills to the tableThe importance of understanding the human side of dataThe fallacy of managing by numbers Chapters 00:00 Data-Driven vs. Data-Inspired Decision-Making02:21 The Journey from Bioinformatics to TED Stage11:47 The Purpose and Limitations of Dashboards19:53 Culture as a Barrier to Data-Inspired Decision-Making28:29 AI and Data Culture29:02 The Role of Data Scientists in the Age of AI30:12 The Misinterpretation of Data34:05 Human Skills in Data Science44:18 The Fallacy of Managing by Numbers

  4. Aug 3

    Your AI Policy Describes a Company That Doesn't Exist

    The conversation delves into AI governance in regulated industries, the gap between what governance policies say and what employees actually do, the human-in-the-loop concept as it works in practice rather than on a process map, and where accountability sits when an AI writes something and a human approves it in four seconds. Daanyaal Bandukwala brings the pharmaceutical commercialization and patient support program lens, and the episode is upfront about the commercial relationship between host and guest, which is why the questioning goes harder than usual. Topics include the Globe and Mail test as a governance heuristic, automation bias in expert decision making, AI disclosure, data residency versus sovereignty, the collapse of the build case in favour of buying, and an honest audit of which governance studies are vendor funded. Takeaways Governance is not killing AI projects. Unanswered ownership questions are. Compliance is asking who is accountable, and nobody wrote it down.The Globe and Mail test is the simplest governance heuristic available. If this ended up on the front page tomorrow, how would you look?AI governance lives in onboarding, not in a paragraph buried in a large SOP.Human in the loop is not the same as human accountability. When AI is wrong, expert accuracy collapses, and only personal accountability changes that.A quarterly spot check beats an annual formal audit, and a vendor's reaction to being spot checked tells you more than the audit clause.No independent, non-vendor study isolates governance as the cause of better AI outcomes. It is conviction, not proof, and both host and guest say so.Start with the business problem, not the platform. Evaluate governance before features.Walk out if a vendor claims 100 percent accuracy, says governance can wait, calls integration phase two, or says AI works the same in any industry.Executives are wrong about how their people feel and wrong about what their people are doing, and most are comfortable with unapproved use anyway.Chapters 00:00 The Role of Governance in AI03:20 Challenges in AI Adoption19:43 Human-in-the-Loop in AI Practice27:55 Accountability in AI30:48 Responsible AI and Transparency31:58 AI Disclosure and Transparency33:13 AI in Medical Education35:18 Influencer Marketing and AI36:37 AI in Healthcare and Patient Support41:20 Data Resiliency and Sovereignty45:12 The Review Step That Stopped Happening50:07 Regulatory Decisions and Drug Discovery53:11 Build vs. Buy in AI Solutions57:48 Specialized AI Vendors58:30 Why There Is No Good Data on Build vs. Buy1:02:11 How to Choose an AI Vendor1:07:16 Vendor Red Flags and When to Walk Out1:11:52 The Hard Question: Is Governance Actually Proven?1:15:28 Five Beliefs, True or False1:19:18 Shadow AI and the Executive Perception Gap1:23:17 The 95 Percent Pilot Failure Myth1:26:57 Whose Job Gets Smaller1:29:18 Middle Management and Job Security1:32:57 The First 90 Days1:36:01 What Daanyaal Is Least Sure About1:39:57 The Guest Question and Closing

  5. Jul 20

    He's Seen This Movie Before

    Mykle McKiernan is an enterprise technology leader who has built and run platforms at the highest level, with senior roles at Wayfair and Amazon and earlier work at McKinsey. He has lived through four enterprise technology revolutions from the inside: spreadsheets, ERP, CRM, and now AI. That vantage point is rare, and it makes him one of the few people who can tell you whether this moment is actually unprecedented or just the newest wave. This conversation traces the impact of AI on enterprise technology by lining it up against the revolutions that came before it. Spreadsheets, ERP, and CRM each arrived with the same fear, the same resistance, and the same eventual payoff, and each one teaches something about what makes technology stick. The throughline is simple: adoption succeeds when people gain leverage and fails when they only see compliance, and value has to reach every stakeholder to hold. From there the discussion turns to what makes this wave different. Spreadsheets democratized analysis, ERP standardized process, CRM standardized relationships, and AI democratizes intelligence itself. That is why the reaction is so much stronger this time. It is no longer about process. It is about identity. The winners will not be the people who compete with AI. They will be the ones who learn to orchestrate it, who understand that real value shows up quietly years later rather than in the demo, and who are willing to let AI into the work and the hobbies they care about most. Takeaways Technology adoption fails when users only see compliance, it succeeds when users gain leverage.Adoption depends on delivering value to multiple stakeholders first and quickly. AI adoption success depends on employee leverage.Stakeholders' mutual benefit is crucial for AI adoption.The AI wave is different because it is about identity, not process.The winners will be those who learn to orchestrate AI.AI value shows up quietly years later, not in the demo.Embrace AI in hobbies for mutual benefit.Chapters 00:00 The Impact of AI on Enterprise Technology02:02 The Spreadsheet Revolution12:46 The ERP Challenge29:48 The CRM Dilemma34:14 AI Adoption and Experimentation37:27 Stakeholders' Mutual Benefit39:29 AI Wave: Identity vs. Process40:13 Orchestrating AI40:58 Value of AI Shows Up Quietly01:05:29 Embracing AI in Hobbies

  6. Jul 13

    Everyone Feels Faster With AI. The Stopwatch Disagrees.

    The conversation delves into the role of AI as a compression engine, its impact on search behavior, and the loss of critical thinking in research. It also explores the semantic search problem and the feature rank system. The conversation delves into the challenges and implications of AI, including its accountability for misinformation, the challenges in AI adoption and ROI, and the importance of understanding the limitations of AI. It also explores the impact of AI on work and the need to understand when not to rely on AI. The discussion highlights the need for a nuanced understanding of AI's capabilities and limitations, as well as the responsibility and accountability associated with its use. Takeaways AI as a compression engineAI's impact on search behaviorThe loss of critical thinking in research AI's accountability for misinformationChallenges in AI adoption and ROIUnderstanding the limitations of AI Chapters 00:00 Semantic Search Problem and Feature Rank System32:37 AI and Misinformation34:05 Accountability and Responsibility35:25 The Challenge of Defining Truth38:31 AI Adoption and ROI43:06 False Expectations and Unattainable Goals50:25 AI's Impact on Work59:16 Understanding AI's Limits

  7. Jul 2

    A 3-Year AI Strategy Is Already Dead

    The conversation explores the era of agentic AI, understanding AI agents and automation, AI agents as workforce augmentation, and scaling and governance of AI agents. It delves into the challenges, risks, and misconceptions surrounding the implementation and management of AI agents in organizations. The conversation covers a range of topics related to AI strategy, agent development, and the use of language models. It delves into the challenges and considerations for leaders in adopting AI technologies and formulating effective strategies. The discussion emphasizes the importance of staying informed, avoiding vendor lock-in, and strategically leveraging language models in the organizational stack. Takeaways AI agents are not meant to replace human workforce but to augment and enhance it.The governance and scaling of AI agents pose significant challenges and risks that need careful consideration and management. Augmented full stack development enables faster agent developmentCaution is needed in buy vs. build approach for agentsAvoiding vendor lock-in and ensuring portability is crucial for long-term AI strategy Chapters 00:00 The Era of Agentic AI06:56 Understanding AI Agents and Automation18:14 Scaling and Governance of AI Agents27:54 Augmented Full Stack Development36:30 Decoupling from Vendor Lock-in45:12 AI Strategy and Leadership Disclaimer: The views and opinions expressed by Brent Lewis in this podcast are his own and do not necessarily reflect the official policy or position of Armstrong World Industries, Inc. or any of its affiliates. Brent's participation in this episode is in his personal capacity, and nothing in this conversation should be construed as an official statement, endorsement, or representation on behalf of Armstrong.

  • 7 Episodes

About

Everyone is talking about AI. Almost nobody is telling the truth about it. This is the show where builders, operators, and the people quietly reshaping enterprises sit down and say what they actually think. No hype cycles. No buzzword bingo. Just honest conversations about what AI is really doing to our companies, our jobs, and the way we make decisions. I'm Shubhendu. I build AI systems for a living, and I've seen what works, what breaks, and what nobody wants to admit in the boardroom. Every week I bring on the people behind the systems, the strategies, and the bets that are defining this decade, and I ask them the questions everyone is thinking but rarely says out loud. If you want the polished keynote version, there are a thousand of those. If you want the real version, pull up a chair.

Information

  • Creator
    Shubhendu Tripathi
  • Years Active
    2026
  • Episodes
    7
  • Rating
    Clean
  • Copyright
    © 2026 Shubhendu Tripathi
  • Show Website
    The Integration Layer