{
  "abstract": "A/B testing remains the standard for rolling out new features in the technology industry. Each experiment, however, consumes real traffic, engineering effort, and weeks of wall-clock time. Can AI agents---conditioned on behavioral profiles and contextual descriptions of the intervention---simulate outcomes accurately enough to vet candidate treatments before committing live traffic? We formalize this question as a \\emph{Simulated Randomized Controlled Trial} (S-RCT) and derive a two-layer error decomposition that separates agent approximation error from subsampling error, enabling targeted improvements to each. The framework is agent-agnostic: any behavioral model---from a fine-tuned specialist to a general-purpose foundation model---can serve as the simulation engine. Validated on 67 historical marketing A/B tests, a baseline S-RCT using an off-the-shelf foundation model captures directional signal (sign overlap 0.70) but systematically overshoots effect magnitudes. A two-phase pre-period calibration protocol reduces the squared prediction error (after removing irreducible measurement noise) by ${\\sim}77\\times$; a within-subject design---where each agent is exposed to both arms---reduces standard errors by ${\\sim}2.4\\times$. We discuss limitations of the current approach and identify applications where experimenters stand to benefit from agentic signals.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2608.02345",
  "author_total": 2,
  "authors": [
    {
      "name": "Stefan Hut"
    },
    {
      "name": "Lorenzo Masoero"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/egyfx856nb4hz.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2608.02345",
  "fulltext_url": null,
  "id": "egyfx856nb4hz",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2608.02345",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "http://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.AI"
    ],
    [
      "arxiv",
      "cs.CL"
    ],
    [
      "arxiv",
      "stat.AP"
    ],
    [
      "openalex",
      "2613|Advanced Causal Inference Techniques"
    ]
  ],
  "openalex": "W7172377494",
  "pdf_url": "https://arxiv.org/pdf/2608.02345",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-11",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/egyfx856nb4hz.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:21:36Z",
      "native_id": "2608.02345",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2608.02345"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:stat.AP",
      "subfield": "statistics"
    }
  ],
  "title": "Can AI Agents Simulate A/B Test Outcomes? A Validation Framework for Agentic Experimentation",
  "title_text": "Can AI Agents Simulate A/B Test Outcomes? A Validation Framework for Agentic Experimentation",
  "type": "preprint",
  "updated_at": "2026-10-30T12:21:36Z",
  "url": "https://research.draconitic.com/paper/egyfx856nb4hz",
  "venue": "arXiv",
  "versions": []
}