The Practical AI Podcast

Taylor
  • 5.0 (1)
  • Arts

The Practical AI Podcast with Hashi Sivananthan, Founder and Principal at Digiform AI doesn't have to be confusing — and it definitely doesn't have to be hype. Every week, AI adoption and governance expert Hashi breaks down what's actually happening in AI and what it means for your business. No jargon. No buzzwords. Just simple, straightforward guidance for non-technical business leaders who want to drive real results with AI — without breaking what already works. From customer experience and personalization to governance, guardrails, and knowing when not to use AI, each episode gives you practical takeaways you can bring to your next leadership meeting. Because good AI starts with good decisions. New episodes weekly. Follow Hashi on Instagram @aimehashi.

Episodes

  1. Oct 14

    Your Customer Asked AI First: Smarter Buyers and the First AI Holiday

    A client brings a two-page contract to a lawyer, gets quoted $10,000 for review, and already knows that's too much. They ran it through ChatGPT before the meeting. In this episode of Practical AI, Hashi Sivananthan, Founder and Principal at Digiform, breaks down how AI turned buyers into researchers across law, accounting, consulting, agencies, and direct-to-consumer. The catch is that much of what AI tells your prospects comes from old Reddit threads, forums, and outdated reviews, repeated as fact. Hashi's position: don't be threatened by it. Manage it. Then the holiday edition. Salesforce predicts about 20% of holiday e-commerce traffic this year will come from AI agents, and Adobe found AI-referred shoppers converted 31% more than other traffic last holiday season. If an AI assistant builds your customer's gift list, does it know you exist? In this episode: How AI closed the knowledge gap between buyers and sellersWhy AI repeats outdated information about your business, and the two fixesThe two problems every brand faces before Black Friday: bot readiness and the shift to AI shoppingA 3-question audit you can run this afternoonHashi's hot take: businesses that can't publish accurate content fast are going to loseSources mentioned: Salesforce, 2026 Holiday Predictions: https://www.salesforce.com/blog/holiday-retail-predictions-2026/Adobe Digital Insights, 2025 holiday AI traffic: https://business.adobe.com/blog/ai-driven-traffic-surges-across-industriesMore from Hashi: The Context Window newsletter: https://ctxwindow.comDigiform: https://digiform.io/

  2. Oct 14

    AI Got Hands: Why Your Machines Talking to Each Other Is the Next Big AI Story

    For three years, AI has lived on a screen. In late August, that changed. Anthropic opened a research preview of the Model Hardware Standard (MHS), a common language that lets AI agents discover, understand, and safely operate physical equipment. That includes microscopes, robotic arms, liquid handlers, and lasers. In this first Practical AI Industry Edition, Hashi explains what MHS actually is and why the boring word “interoperability” is the whole story. He also covers what this looks like in a real lab today, and what happens when it reaches factory lines and doctors’ offices. In this episode: What Anthropic released, in plain English, and why it isn’t locked to one modelInteroperability, defined simply: can your machines talk to each other?The smart-home lesson: a promised connected home that delivered a drawer full of appsEarly partner results from Carnegie Mellon and QuEra, per AnthropicFactory quality control and the eye exam that summarizes itselfWhat this means for hands-on jobs: help, not replacementChapters: 0:00 AI got hands0:29 What we’re covering0:49 What Anthropic released: the Model Hardware Standard1:51 Interoperability in plain English2:43 Inside a lab experiment, before and after3:32 Early results from Carnegie Mellon and QuEra4:29 Outside the lab: factories and doctors’ offices5:10 What this means for jobs5:37 Practical takeaway6:00 Hot takePractical takeaway: List every place in your operation where a person’s job is carrying information from one machine to another. That’s where this lands first. The next time you buy equipment, ask your vendor one question: does it have an API? Hot take: The next big AI story isn’t a smarter chatbot. It’s your equipment finally talking to each other, and whoever connects their machines first will run circles around everyone else. Next episode: from the lab to your checkout page. Your customers are asking AI about you before they ever talk to you, and the holidays are about to make that obvious.

  3. Oct 14

    The AI Risk Already Inside Your Company

    A researcher quit one of the biggest AI labs in the world so he could say it out loud: the people building AI believe it could wipe us out. Then one of that lab's own alignment leads publicly agreed. So should you stop using AI at work on Monday? Short answer: no. In this episode, Hashi breaks down what actually happened when Jacob Coxon resigned from Anthropic, why the story became headline news, and what the panic skipped. The warning itself isn't new. What's new is who's saying it: not critics outside the industry, but the builders. Then Hashi takes apart the question everyone asks, "Is AI becoming sentient?", and replaces it with the one that matters. Sentient means it has feelings. Aligned means it does what we intended. A system doesn't need to feel anything to cause harm. It only needs a badly described goal and enough access to chase it. Tell an agent to cut costs and it might cancel your insurance. It didn't hate you. It did what you said, not what you meant. For balance, the episode covers RAND's analysis of how AI could actually cause extinction, and why researchers found it immensely challenging but couldn't rule it out. Then the business lens: the doomsday scenario belongs to labs and governments. The alignment problem is already in your building, wherever AI agents are connected to email, calendars, payments and customer data. Chapters 00:00 Cold open: the builders are worried00:18 What we're covering00:50 What actually happened01:32 Why this time is different02:08 "Is AI sentient?" is the wrong question02:32 The cut-our-costs example03:09 What RAND found03:35 The alignment problem in your building03:59 Three guardrails for this quarter04:33 What not to do: don't freeze04:48 Practical takeaway05:08 Hashi's hot take05:22 Next episodePractical takeaway Stop asking whether AI will end the world. Ask what the AI in your company can do without a human saying yes. If you can't answer that in one sentence, that's your first project this quarter. Three guardrails get you started: a human approves anything irreversible, every agent gets least access, and someone owns the inventory of what AI is running and what it can touch. Hot take The apocalypse debate is a distraction for most businesses. The AI that hurts you won't be a superintelligence. It'll be the agent you gave admin access and forgot about. Chapter times are estimated from the VEED transcript; confirm against the final edit, especially if the cold open is re-recorded. The takeaway uses "can't" pending the 04:56 audio check.

  4. Aug 30

    What If You Never Had to Remember Anything Again?

    What if you never had to remember anything again? No to-do lists, no "don't forget the milk," no 3 a.m. panic about the appointment you didn't book. That's not a productivity hack — it's the personal language model, and if Hashi could snap his fingers and raise the funding, it's the company he'd start. In this episode, Hashi picks up where the single-serving AI story left off: single-purpose devices are winning, so the next move is those devices talking to each other. That connection is the starting point of the personal language model — an AI that learns your patterns, gathers your to-dos, and pushes them to you proactively at the exact moment you're most receptive. Today's assistants are learning about you but can't really act on it yet. Once AI can take action on your behalf, the gateway opens. Then the obvious objection, taken seriously: if an AI remembers everything for you, do you get dumber? Hashi's answer is the inverse. You won't outsource knowledge — you'll outsource the mundane. The doctor's appointment, the license renewal, the milk. The anxiety of holding all that eats real cognitive capacity, and freeing your memory doesn't shrink your mind; it gives it room. The episode also puts interoperability in plain English — can your devices talk to each other? — with the smart-home flashback as the cautionary tale: consumers were promised a connected home and got a junk drawer full of apps. Expect ecosystems instead, and one non-negotiable gate for every product that wants into your personal space: no trust, no personal AI. In this episode The personal language model, defined — and why it's the company Hashi would startFrom single-purpose devices to devices that share what they learnProactive AI: to-dos gathered and delivered when you're actually receptive"Won't this make us dumber?" — the secondary-memory answerInteroperability in plain English, and the smart-home junk drawerEcosystems, hardware-plus-model companies, and the privacy gateThe smart fridge, Whirlpool, and why consumer behavioral shifts belong in your product forecast todayPractical takeaway If you're a consumer: expect AI embedded in devices you already own, doing one job really well — and judge every product by how it treats your data. If you're a business owner: start mapping the consumer behavioral shifts now, walk them back into your product planning, and treat privacy and security as the feature — that's what earns the right to be in someone's personal space. Hashi's hot take AI in your personal life: good, one hundred percent. It won't make us stupid — it'll do the inverse. Taking the mundane off our shoulders means more freedom, more presence, probably better mental health. If you think outsourcing your to-do list rots your brain, so did writing it on paper. Practical AI helps non-technical business leaders understand AI and technology in a way that's hype-free and actually relevant to real situations. New episodes covering what AI means for your business, without the noise.

  5. Aug 24

    When Everyone Has AI, What's Your Edge?

    Right now, every company is scrambling to get AI in the door. Very soon, every company will have it — same tools, same automation, same productivity gains. And when the gains are equal, they stop being an advantage. In this episode, Hashi starts with what AI is actually doing to jobs right now — not the headlines, the reality. It isn't replacing the people already in the workplace. It's delaying hiring in roles built on repetitive manual work, and it's consolidating roles: the UI/UX designer expands into graphic design, the backend developer becomes full stack. If AI can help someone with zero coding experience ship an app, imagine what an engineer with ten years of practice builds with the same tools. Then the through line. Everyone is racing to say "yes, we have AI." Almost nobody is asking what comes after — the day every company's AI is practical, running, and efficient, and the playing field levels. Hashi talks to executives every week, and that question barely comes up. His answer: when the tools are the same everywhere, the only things AI can't level are your relationships and your data. Nobody else has your customers' trust and history, and nobody else has the data your business generates every day. In this episode What AI is actually doing to jobs right now — delayed hiring, not replacementRole consolidation: why one person now covers what used to take twoThe frenzy problem: everyone solving today, nobody planning for afterWhat happens when every company has the same AIThe two assets that can't be leveled: relationships and dataA preview of personal AI at work — where this series goes nextPractical takeaway Don't stop at getting AI running — everyone will get there. Start investing now in the two things the level playing field can't touch: deepen the customer relationships only you hold, and treat the data your business generates every day as the asset it's about to become. Hashi's hot take Your AI strategy won't save you. In a few years, every business will have the same AI. The only edges left are your relationships and your data — and if you're not investing in those right now, you're building on rented ground. Chapters 00:00 Everyone is getting AI — then what? 00:17 When everyone has AI, what's your edge? 00:25 What we're covering 00:41 AI isn't replacing workers — it's delaying hiring 00:57 Role consolidation: one person, two jobs 01:24 The frenzy, and the day the gains equalize 01:59 The two edges AI can't level 02:18 Preview: personal AI at work 03:09 Hashi's hot take

  6. Aug 19

    The Do-Everything AI Gadget Is Dead: What will Prevail

    Heavily funded, beautifully launched, and largely gone. The Rabbit R1 was supposed to be the device that let AI solve everything, and its failure tells us more about where AI hardware is heading than any of the successes do. In this episode, Hashi breaks down why the do-everything AI gadget flopped — not because the AI was weak, but because "everything" is a terrible product spec. A device trying to be your assistant, your phone, your camera and your therapist ends up a worse version of all four. Meanwhile the AI devices people actually adopted had one thing in common: they did a single job exceptionally well. That pattern has a name. Single-serving AI. One device, one purpose, executed so cleanly you forget there's a model behind it. A pillbox that notices you're two days from running out and has the refill waiting at Walgreens before the thought occurs to you. Kitchen, bathroom and bedroom devices quietly handling their one job without being asked. And then the part that matters most: what happens when those devices start talking to each other. That connection layer is where this stops being convenient and starts being transformative — and it's the thread running into the next few episodes. In this episode Why heavily funded do-everything devices didn't deliverWhat the AI devices people actually kept had in commonSingle-serving AI, defined — and why it beats the super gadgetThe pillbox, the kitchen, the bathroom: what this looks like in practiceWhy devices talking to each other is the real frontierThe privacy question you should be asking now, not laterPractical takeaway If you're building or buying AI hardware, stop chasing the everything device. Look for AI embedded in the devices you already use, focused on doing one thing well — and start asking the second question now: how will these devices talk to each other without putting my privacy and my data at risk? Hashi's hot take The do-everything AI gadget is dead. One device that does everything will always lose to ten devices that each do one thing perfectly. Chapters 00:00 Welcome 00:26 Remember the Rabbit R1? 00:44 What we're covering 00:53 Why the do-everything gadgets flopped 01:26 What the winners had in common 01:43 Single-serving AI, defined 01:52 Imagine a world where your pillbox reorders itself 02:30 What happens when devices talk to each other 02:51 The practical takeaway 03:17 Hashi's hot take Practical AI helps non-technical business leaders understand AI and technology in a way that's hype-free and actually relevant to real situations. New episodes covering what AI means for your business, without the noise.

  7. Jul 15

    Is Generative UI Helping or Hurting Your Customer Experience?

    Generative UI promises every customer a personalized, dynamically-generated interface — but is it actually improving the experience, or quietly breaking it? In this episode, Hashi cuts through the AI hype to explain what Generative UI (GenUI) really is, why companies are racing to adopt it, and where it goes wrong. From e-commerce checkouts to the Marriott app, he explores the real promise of AI-personalized interfaces — and the "unwritten contract" between a user and a button that too many companies are breaking. You'll learn: What Generative UI is, in plain business languageWhy personalization drives conversion — and why habit beats personalizationWhat a 2026 Qualtrics report reveals about the personalization-privacy tradeoffThe #1 mistake companies make when deploying GenUIA practical framework for rolling out Generative UI safely: guardrails, data readiness, and controlled experimentsIf you're a non-technical business leader trying to figure out where AI actually fits in your customer experience strategy, this one's for you. Follow Hashi on Instagram: @aimehashi

  • 7 Episodes

About

The Practical AI Podcast with Hashi Sivananthan, Founder and Principal at Digiform AI doesn't have to be confusing — and it definitely doesn't have to be hype. Every week, AI adoption and governance expert Hashi breaks down what's actually happening in AI and what it means for your business. No jargon. No buzzwords. Just simple, straightforward guidance for non-technical business leaders who want to drive real results with AI — without breaking what already works. From customer experience and personalization to governance, guardrails, and knowing when not to use AI, each episode gives you practical takeaways you can bring to your next leadership meeting. Because good AI starts with good decisions. New episodes weekly. Follow Hashi on Instagram @aimehashi.

Information

  • Creator
    Taylor
  • Years Active
    2026
  • Episodes
    7
  • Rating
    Clean
  • Copyright
    © 2026 The Practical AI Podcast
  • Show Website
    The Practical AI Podcast