{
  "video": {
    "id": "C2-v6wkHXhY",
    "title": "Your AI Isn't the Advantage. Your Judgment Is - Sid Pednekar",
    "duration": 1093,
    "upload_date": null,
    "channel": "AI Engineer",
    "source": "AI Engineer",
    "event_format": ""
  },
  "analysis": {
    "video_id": "C2-v6wkHXhY",
    "title": "Your AI Isn't the Advantage. Your Judgment Is - Sid Pednekar",
    "one_liner": "When AI makes intelligence cheap and abundant, judgment becomes the scarce resource, so leading organizations ask \"where is our best judgment worth spending?\" and redesign work around the answer instead of adding AI to existing workflows.",
    "summary": "Sid argues that AI access is widespread but its impact is uneven, citing studies in which only about 37% of surveyed enterprises report a positive EBIT contribution and about 6% qualify as high performers. He says leaders stand out through three pillars: designing the right problem, building shared capabilities, and compounding the advantage, all tied together by judgment as a finite asset. Two hypothetical asset-management firms illustrate the gap: company B uses a co-pilot to summarize earnings transcripts faster, while company A asks why material signals reach it too late and redesigns the analyst's work around an investment thesis.",
    "key_points": [
      "Cited studies show only about 37% of surveyed enterprises reporting a positive EBIT contribution, with about 6% qualifying as high performers (about 5% contribution to EBIT). Other studies show about 66% reporting productivity increases, but only about 40% and 20% seeing cost reductions and revenue growth.",
      "Sid defines judgment as choosing what matters and how to act when there is no single obvious answer. It combines creativity (which possibilities are worth exploring), taste (what good looks like in context) and decisions (which trade-offs are worth making).",
      "Starting with the tool produces a \"faster horse\": company B feeds earnings transcripts through a summarization workflow and loses context through information reduction. Company A asks why material signals reach it too late and where its bottleneck is.",
      "Leaders concentrate on fewer, higher-quality decisions and spend more than 80% of their AI investment on high-value use cases that transform key business functions or create new offerings. Scattered pilots risk \"innovation theater\" or \"productivity theater\" rather than actual value generation.",
      "Winners redesign rather than just automate. In company A's hypothetical workflow, the analyst defines the investment thesis, gathers evidence that supports or refutes it, monitors continuously and investigates exceptions. Sid cites high performers redesigning workflows at 75% versus 25% for the rest.",
      "Technology and business are converging: engineers are moving closer to business outcomes, business professionals are building tools and automations, and forward deployed portfolio managers are possible alongside forward deployed engineers and product managers. More large enterprises see their technology arm as a business growth function.",
      "\"Standardize the core and personalize the edges\": instead of one-off workflows, data scientists could share a core of data sets, semantic layers, trade-offs, what worked and known gaps, so each new person becomes almost a plug-and-play module.",
      "Enterprises were built for humans interacting with a single application, not for agents working across workflows and applications. Sid cites 71% of orgs modernizing core infrastructure for AI and 23% investing 6 to 10% of annual revenue in core system modernization, with data quality and reuse as common challenges. He describes three types of team agents: expert agents, workflow agents (such as Slack bots that open and close tickets) and chief-of-staff agents that track the state of the organization."
    ],
    "takeaways": [
      "Before choosing a tool, ask what the right problem is and where the bottleneck sits, for example \"Why are our material signals reaching us too late?\" rather than \"Where can we add AI to the analyst work?\"",
      "Concentrate investment on end-to-end, high-value workflows, aligning engineering, analytics, AI capacity and judgment on a singular focus, and separate experimentation from actual value generation.",
      "Redesign roles and processes around where human judgment is needed instead of speeding up each step of the existing workflow.",
      "Break down organizational silos and build shared cores of reusable knowledge such as data sets, semantic layers and records of what worked, asking what the next person should not have to rebuild.",
      "Feed corrections and successes back into the system through a new tool or context, preserve the evidence with those learnings, and direct capacity freed by AI toward key metrics you evaluate and monitor to create a flywheel."
    ],
    "topics": [
      "ai strategy",
      "judgment",
      "workflow redesign",
      "enterprise adoption",
      "asset management",
      "shared capabilities",
      "agents",
      "organizational learning"
    ],
    "tools": [
      "Slack"
    ],
    "quotes": [
      {
        "text": "more answers don't necessarily create more value.",
        "at": "04:03",
        "url": "https://www.youtube.com/watch?v=C2-v6wkHXhY&t=243s"
      },
      {
        "text": "where is our best judgment worth spending?",
        "at": "05:20",
        "url": "https://www.youtube.com/watch?v=C2-v6wkHXhY&t=320s"
      },
      {
        "text": "So, the winners just don't just automate, they redesign.",
        "at": "11:09",
        "url": "https://www.youtube.com/watch?v=C2-v6wkHXhY&t=669s"
      },
      {
        "text": "So, what should the next person not have to rebuild?",
        "at": "14:31",
        "url": "https://www.youtube.com/watch?v=C2-v6wkHXhY&t=871s"
      }
    ],
    "words": 2942
  },
  "summary_url": "/#C2-v6wkHXhY",
  "transcript": {
    "html": "/transcripts/C2-v6wkHXhY.html",
    "txt": "/transcripts/C2-v6wkHXhY.txt",
    "vtt": "/transcripts/C2-v6wkHXhY.vtt"
  }
}