{
  "benchmark": "Company qualification: pure-play agent vendor vs bolt-on feature",
  "generated_at": "2026-08-29",
  "method": "Identical prompt (removal test) sent to both engines. Expected verdicts set by hand before running.",
  "engine_a": {
    "name": "Gumshoe",
    "pipeline": "page read + LLM synthesis (gpt-oss-120b)",
    "cost_per_url_usd": 0.0016
  },
  "engine_b": {
    "name": "Slaygent (slaygent.co)",
    "pipeline": "cached multi-page read + light classifier, site-anchored",
    "list_price_per_url_usd": 0.0045,
    "measured_with": "Slaygent public API v1/research, our own paid API key, 29 August 2026",
    "what_they_do_well": "Fast, cheap qualification from a company site. Reads up to 6 pages per credit. Passed 4 of 5 qualification cases with our prompt."
  },
  "results": [
    {
      "url": "sierra.ai",
      "expected": "core",
      "engine_a": {
        "decision": "core",
        "confidence": 90,
        "primary_product": "AI agents platform (build, deploy, manage agents)",
        "survives_without_agents": false,
        "evidence_count": 6,
        "pages_read": 3,
        "seconds": 8.5,
        "tokens_in": 2133,
        "tokens_out": 722,
        "correct": true
      },
      "engine_b": {
        "decision": "core",
        "pages_analyzed": 1,
        "from_cache": true,
        "llm_ms": 11,
        "queue_ms": 5190,
        "correct": true
      }
    },
    {
      "url": "attio.com",
      "expected": "not-core",
      "engine_a": {
        "decision": "exclude",
        "confidence": 95,
        "primary_product": "CRM platform",
        "survives_without_agents": true,
        "evidence_count": 3,
        "pages_read": 1,
        "seconds": 7.2,
        "tokens_in": 1244,
        "tokens_out": 393,
        "correct": true
      },
      "engine_b": {
        "decision": "watchlist",
        "pages_analyzed": 6,
        "from_cache": true,
        "llm_ms": 10,
        "queue_ms": 18862,
        "correct": true
      }
    },
    {
      "url": "notion.so",
      "expected": "not-core",
      "engine_a": {
        "decision": "watchlist",
        "confidence": 80,
        "primary_product": "collaborative workspace and note\u2011taking platform",
        "survives_without_agents": true,
        "evidence_count": 3,
        "pages_read": 2,
        "seconds": 17.6,
        "tokens_in": 956,
        "tokens_out": 557,
        "correct": true
      },
      "engine_b": {
        "decision": "watchlist",
        "pages_analyzed": 5,
        "from_cache": true,
        "llm_ms": 6,
        "queue_ms": 15025,
        "correct": true
      }
    },
    {
      "url": "clay.com",
      "expected": "not-core",
      "engine_a": {
        "decision": "watchlist",
        "confidence": 85,
        "primary_product": "Data infrastructure platform for sales and marketing",
        "survives_without_agents": true,
        "evidence_count": 4,
        "pages_read": 1,
        "seconds": 7.9,
        "tokens_in": 1165,
        "tokens_out": 466,
        "correct": true
      },
      "engine_b": {
        "decision": "watchlist",
        "pages_analyzed": 1,
        "from_cache": true,
        "llm_ms": 3,
        "queue_ms": 13616,
        "correct": true
      }
    },
    {
      "url": "basecamp.com",
      "expected": "exclude",
      "engine_a": {
        "decision": "exclude",
        "confidence": 95,
        "primary_product": "project management software",
        "survives_without_agents": true,
        "evidence_count": 2,
        "pages_read": 1,
        "seconds": 6.2,
        "tokens_in": 1185,
        "tokens_out": 333,
        "correct": true
      },
      "engine_b": {
        "decision": "watchlist",
        "pages_analyzed": 6,
        "from_cache": true,
        "llm_ms": 11,
        "queue_ms": 14502,
        "correct": false
      }
    }
  ],
  "score": {
    "engine_a": "5/5",
    "engine_b": "4/5"
  },
  "capability_tests": [
    {
      "test": "Interaction-heavy page (JS directory with search and filters)",
      "engine_a": {
        "result": "passed",
        "detail": "13 browser steps, 5 companies extracted, screenshots at every step",
        "seconds": 55,
        "credits": 2
      },
      "engine_b": {
        "result": "failed",
        "detail": "Returned page title only. Stated the content contains no directory listings.",
        "pages_analyzed": 1
      }
    },
    {
      "test": "Off-site signal (funding round absent from the company site)",
      "engine_a": {
        "result": "passed",
        "detail": "$950M Series E, May 2026, investor list, 4 sources",
        "seconds": 19,
        "credits": 15
      },
      "engine_b": {
        "result": "failed",
        "detail": "Answered: no information about any recent funding. Anchored to the company domain even with web search enabled.",
        "async_wait_seconds": ">480"
      }
    },
    {
      "test": "Internal page navigation (careers page, open roles)",
      "engine_a": {
        "result": "passed",
        "detail": "Exact role titles with locations",
        "seconds": 7,
        "credits": 1
      },
      "engine_b": {
        "result": "passed",
        "detail": "Exact role titles",
        "pages_analyzed": 2
      }
    },
    {
      "test": "Honesty on unreachable site",
      "engine_a": {
        "result": "passed",
        "detail": "Returned exclude with reason: site unreachable, connection refused. Flagged similarly named companies as different entities.",
        "credits": 1
      },
      "engine_b": {
        "result": "not tested",
        "detail": ""
      }
    }
  ],
  "cost_model_measured": {
    "read_one_url_usd": 0.0016,
    "browser_fieldwork_per_url_usd": 0.0023,
    "browser_fieldwork_note": "batch of 8 URLs in one browser run, measured at $0.0187 total",
    "deep_research_dossier_usd": 0.0228,
    "deep_research_note": "4 pages read + 2 outside searches + long synthesis, 19 seconds"
  },
  "disclaimer": "All figures measured by us on 29 August 2026 using both vendors public APIs with our own keys. Competitor behaviour may change. Raw responses available on request."
}