{
  "video_id": "KVFrBtkLg4I",
  "channel_slug": "ibmtechnology",
  "channel_handle": "IBM Technology",
  "title": "How Agentic AI Transforms Maintenance and Asset Decisions",
  "duration_seconds": 311,
  "url": "https://www.youtube.com/watch?v=KVFrBtkLg4I",
  "upload_date": "20260519",
  "transcript": "Anytime you buy something, you want it to operate reliably and last.\nA house, an appliance, a car,\nbut businesses experience the same thing with their assets,\nthink of a bridge,\nan airplane,\nor even a production plant.\nUnplanned outages and breakdowns can cost hundreds and thousands of dollars per hour, if not maintained.\nSo how does agentic AI impact these asset-intensive industries and help prevent very real, very expensive problems?\nFor decades, we've managed these assets using systems of record.\nThey track data related to assets, operations, and management, such as...\nAsset details, work orders, and inventory.\nThey tell us things like, what has changed, when it changed,\nqnd who changed it.\nThe data is continuously captured in a variety of ways, synthesized, planned, and executed.\nBut the real challenge is turning that data into the right actions while balancing the trade-off decisions.\nAutomated workflows help, but too many of those decisions still depend on too few skilled people.\nAnd that doesn't scale.\nSo systems have to evolve.\nAnd that's where agentic AI comes in.\nWe're seeing a shift from systems of record to systems of intelligent action.\nThis doesn't replace the record.\nIt runs on top of it.\nIt reasons,\nit plans,\nand it acts.\nTogether, this is where a genetic AI takes us beyond analysis into systems that act with purpose and operational context.\nLet's imagine a technician is scheduled for a complex repair.\nIn a standard system of record flow,\na person manually prepares the work order, schedules it, and assigns it to a technician.\nIn an intelligent system of action flow, an AI agent.\nDoes the heavy lifting before anyone even logs in.\nThen it provides the work order to our maintenance manager who approves.\nOnce approved, the technician logs in to find a work order that's already scheduled with parts, tools, and diagnostic guidance.\nBut intelligence doesn't stop at planning.\nLet's follow that technician into the field.\nNow the technician is on site looking at existing data.\nThe AI agent determined a root cause based on sensor data and the graded performance of a pump.\nFrom here, the tech and the agent work together hand in hand in every stage.\nThe technician works hands-free, describing what they observe verbally, unusual vibration, a visible leak.\nThe technician can also use the camera on a mobile device or smart glasses to capture what they see.\nThe agent processes that visual input and overlays procedural guidance in real time.\nHelping diagnose and repair the problem on the spot.\nAn intelligent system of action doesn't stop at advice.\nToday, incomplete closeouts are one of the biggest sources of rework and compliance gaps.\nCritical steps get skipped.\nDocumentation gets deferred.\nParts go unrecorded.\nIn an intelligent system of action, it will catch what's often missed.\nIt prompts in real time to ensure Documentation is captured.\nCompliance steps are completed.\nParts used are recorded, and follow-up inspections are scheduled.\nThe work isn't done when the repair is done.\nIt's done when record is complete.\nFor decades, enterprise software recorded the past.\nNow it can reason about the future, from systems of record to intelligent systems of action, powered by agentic AI.\nThank you.",
  "transcript_chars": 3249,
  "ingested_at": "2026-05-21T19:13:24.464794+00:00",
  "source": "retry-no-transcript",
  "yt_meta": {
    "view_count": 5052,
    "like_count": 175,
    "channel_id": "UCKWaEZ-_VweaEx1j62do_vQ"
  }
}