{
  "schema_version": 1,
  "builder_plan": {
    "scenario": "An agent looks good in demos, but you cannot tell which product tasks it fails.",
    "change": "Use the study's evaluation idea to inspect a small set of tasks with human-checked answers.",
    "baseline": "Your current evaluator and a manually reviewed reference set.",
    "steps": [
      "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.",
      "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.",
      "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.",
      "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test."
    ],
    "measure": "Agreement with human review, missed failures, false alarms, and evaluation cost per task.",
    "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation."
  },
  "builder_review": null,
  "arxiv_id": "2610.23673",
  "title": "On the possibilities and limitations of LLM-based multi-agent systems for multi-criteria decision-making in healthcare operations management",
  "summary": "The framework replaces the unaudited use of LLM-generated stakeholder weights as inputs to multi-objective optimization with an explicit reliability assessment plus robust weighted-sum optimization under preference uncertainty. It costs additional inference calls for eliciting explicit and implicit preferences across roles and conditions, statistical modeling of pairwise choices, and solving robust rather than nominal optimization models, with no claim of reduced compute or latency. It can fail because deliberation among agents erases legitimate role differences, pairwise choices are dominated by presentation order rather than context, and contextual predictive models add little over order alone, so the elicited preferences may not reflect real stakeholder trade-offs even after auditing.",
  "rationale": "The paper is a cautionary evaluation study for a niche healthcare decision-support setting and offers the reader no deployable efficiency or quality gain, so it is relevant only as methodological awareness rather than production adoption.",
  "technique": "Auditing LLM-agent preference weights for decision-making",
  "family": "avaliacao_de_agente",
  "recommendation": "observar",
  "published": "2026-10-16",
  "as_of": "2026-10-19",
  "editorial_status": "indexed",
  "source_url": "https://arxiv.org/abs/2610.23673",
  "independent_implementations": 0,
  "total_implementations": 1,
  "stars": 10,
  "citations": null,
  "claims": [
    {
      "claim_id": "claim-01",
      "statement": "The framework replaces the unaudited use of LLM-generated stakeholder weights as inputs to multi-objective optimization with an explicit reliability assessment plus robust weighted-sum optimization under preference uncertainty. It costs additional inference calls for eliciting explicit and implicit preferences across roles and conditions, statistical modeling of pairwise choices, and solving robust rather than nominal optimization models, with no claim of reduced compute or latency. It can fail because deliberation among agents erases legitimate role differences, pairwise choices are dominated by presentation order rather than context, and contextual predictive models add little over order alone, so the elicited preferences may not reflect real stakeholder trade-offs even after auditing.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.23673",
      "source_page": null,
      "source_excerpt": ""
    }
  ],
  "exposure_map": [
    {
      "dimension": "quality",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "compute",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "latency",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "operations",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "compatibility",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "security",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "data_and_training",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "reproducibility",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    }
  ],
  "risks": [],
  "minimum_test": [],
  "open_questions": [],
  "independent_tests": [],
  "lab_reproductions": [],
  "report_available": false,
  "validation_tier": "unknown",
  "evidence_tier": "unknown",
  "infrastructure_basis": "unknown",
  "equations": [],
  "equations_status": "not_fetched",
  "core_kind": "",
  "equations_fetched_at": "",
  "equations_source_url": "",
  "equations_source_sha256": "",
  "exposures_reviewed_at": "",
  "exposures_source_sha256": ""
}