{
  "$schema": "https://peakstate.global/sourced/v1.schema.json",
  "sourced": "1.0",
  "artefact": {
    "path": "src/content/posts/leading-leaders-agents.md",
    "slug": "leading-leaders-agents",
    "sha256": "176cf69b5ec4caa6b3be358c9b5c79b0dd6dccf97136d6d39b79c4a20b7476c1",
    "producedAt": "2026-08-26"
  },
  "claims": [
    {
      "id": "c1",
      "statement": "A manager who used to run a team of eight now runs three people who each direct a fleet of AI agents; output is up and headcount is flat.",
      "status": "inferred",
      "note": "Composite opening scene drawn from Andrew's coaching conversations. Not one named client and not a case study.",
      "locator": {
        "section": "opening"
      },
      "evidence": []
    },
    {
      "id": "c2",
      "statement": "When capable agents join a team, four things change at once for the first-line manager: allocation becomes designing the human/machine split, quality control becomes designing the checking regime, development gets harder because agents absorbed the junior tasks, and accountability for agent-produced error becomes ambiguous.",
      "status": "inferred",
      "note": "Andrew's own synthesis from client work. No study retrieved.",
      "locator": {
        "section": "What has already changed one level down"
      },
      "evidence": []
    },
    {
      "id": "c3",
      "statement": "These managers are measured with instruments built for the old job (utilisation, headcount, activity), and the gap produces a persistent low-grade stress that shows up in coaching.",
      "status": "inferred",
      "note": "First-person practitioner observation from Andrew's coaching practice. Not measured.",
      "locator": {
        "section": "What has already changed one level down"
      },
      "evidence": []
    },
    {
      "id": "c4",
      "statement": "Leading leaders of agent-assisted teams requires four shifts: review the system rather than the output, count judgement rather than capacity, name a human owner for every consequential output, and lead people through the repricing of their expertise.",
      "status": "inferred",
      "note": "Andrew's prescription. Untested against any control or comparison group.",
      "locator": {
        "section": "What the job above has to become"
      },
      "evidence": []
    },
    {
      "id": "c5",
      "statement": "Organisational capacity has long been arithmetic: people times hours times skill.",
      "status": "recalled",
      "note": "General management commonplace, stated from memory. No source retrieved.",
      "locator": {
        "section": "You count judgement, not capacity"
      },
      "evidence": []
    },
    {
      "id": "c6",
      "statement": "The ratio of agent-hours to human-hours keeps tipping towards agents in exposed knowledge functions, faster than planning cycles allow for.",
      "status": "recalled",
      "note": "UNGROUNDED. The original draft asserted 'every trend line says the ratio inverts within a few years'. No trend data was retrieved, so the rewrite states it explicitly in the body as Andrew's own read and carries no figure and no timeframe.",
      "locator": {
        "section": "Where I see this going wrong"
      },
      "evidence": []
    },
    {
      "id": "c7",
      "statement": "Three failure patterns recur: planning as though today's human/agent ratio is permanent, flattening structure because 'the agents do the work', and leaving the operating-model redesign to the middle layer.",
      "status": "inferred",
      "note": "Pattern observed across a small number of client organisations. Not a sample and not counted.",
      "locator": {
        "section": "Where I see this going wrong"
      },
      "evidence": []
    },
    {
      "id": "c8",
      "statement": "The junior work agents absorbed was also the training ground where senior judgement used to grow, so organisations are living off a stock of judgement they are not replacing.",
      "status": "inferred",
      "note": "Andrew's conclusion, drawn from c2. Plausible and unmeasured; the falsifier is any evidence that senior judgement forms as well without an apprenticeship period.",
      "locator": {
        "section": "What you can do this quarter"
      },
      "evidence": []
    }
  ],
  "evidence": [],
  "disclosure": {
    "attribution": "Written by Andrew Ramsden. AI tools assisted research and drafting; all outputs verified.",
    "accountable": "Andrew Ramsden.",
    "limitations": "No external sources.",
    "references": "No external sources. Claim labels are in the [SOURCED sidecar](/sourced/leading-leaders-agents.sourced.json)."
  }
}