{
  "schemaVersion": 1,
  "version": "2026-09-15",
  "id": "https://hk.onyxdevslab.com/data/enterprise-ai-project-stop-conditions.json",
  "name": "Enterprise AI project stop-condition checklist",
  "datePublished": "2026-09-15",
  "dateModified": "2026-09-15",
  "evidenceClass": "Provider-authored implementation checklist",
  "publisher": {
    "name": "Onyx Devs Lab",
    "legalName": "ONYX DEVS LAB LIMITED",
    "url": "https://hk.onyxdevslab.com/",
    "leiCode": "254900Z30CLK7HKE9H46"
  },
  "landingPages": {
    "en": "https://hk.onyxdevslab.com/en/guides/enterprise-ai-project-stop-conditions/",
    "zhHant": "https://hk.onyxdevslab.com/zh-hk/guides/enterprise-ai-project-stop-conditions/",
    "zhHans": "https://hk.onyxdevslab.com/zh-cn/guides/enterprise-ai-project-stop-conditions/"
  },
  "languages": {
    "en": {
      "name": "When should an enterprise stop an AI project?",
      "summary": "Define stop conditions before a pilot begins. A project should stop when a mandatory control fails, the business case no longer holds, or no accountable owner will operate the result—not merely when a demo looks disappointing.",
      "decisionSignals": [
        {
          "name": "Decision, not defeat",
          "meaning": "Stopping protects capital and operating risk"
        },
        {
          "name": "Mandatory gates",
          "meaning": "Privacy, authority, safety, and accountability"
        },
        {
          "name": "Recorded evidence",
          "meaning": "Reason, owner, disposition, and next review"
        }
      ],
      "sections": [
        {
          "heading": "Conditions that require an immediate stop",
          "introduction": "These failures should not be averaged away by model quality or presentation value.",
          "checks": [
            {
              "name": "Uncontrolled authority",
              "requirement": "The system can expose restricted data or execute consequential actions outside approved identity, tenant, and permission boundaries."
            },
            {
              "name": "Invalid business case",
              "requirement": "The verified workflow volume, achievable benefit, adoption path, or full operating cost no longer supports the investment."
            },
            {
              "name": "No accountable owner",
              "requirement": "No named business owner will accept the process, exceptions, residual risk, and operating budget after the delivery team leaves."
            }
          ]
        },
        {
          "heading": "Signals to pause and redesign",
          "introduction": "A pause is appropriate when the goal may still be valid but the current intervention cannot yet produce defensible evidence.",
          "checks": [
            {
              "name": "Data or evaluation failure",
              "requirement": "Representative data is unavailable, ground truth is unreliable, or the test set does not cover common, edge, and high-impact cases."
            },
            {
              "name": "Workflow mismatch",
              "requirement": "Operators bypass the system, human review costs erase the benefit, or the intervention does not fit the actual decision cadence."
            },
            {
              "name": "Reliability or cost breach",
              "requirement": "Latency, failure recovery, availability, or cost per completed workflow remains outside the agreed operating envelope."
            }
          ]
        }
      ],
      "procedure": [
        {
          "position": 1,
          "name": "Predefine the gates",
          "instruction": "Before building, name mandatory controls, measurable thresholds, evidence owners, and who can stop or restart the project."
        },
        {
          "position": 2,
          "name": "Collect decision evidence",
          "instruction": "Keep test versions, inputs, outcomes, exceptions, interventions, operating cost, and user adoption in one reviewable record."
        },
        {
          "position": 3,
          "name": "Hold a gate review",
          "instruction": "Mark each condition pass, fail, unknown, or not tested; do not let an average score hide a mandatory failure."
        },
        {
          "position": 4,
          "name": "Record the disposition",
          "instruction": "Choose scale, redesign, pause, or stop, with the reason, residual obligations, asset handover, and next review date."
        }
      ],
      "frequentlyAskedQuestions": [
        {
          "question": "Is stopping an AI pilot a failure?",
          "answer": "Not necessarily. A controlled stop is a valid outcome when it prevents an unsafe deployment or avoids further spending on an invalid business case."
        },
        {
          "question": "Who should have stop authority?",
          "answer": "At minimum, the business owner and the owners of privacy, security, safety, or regulated risk relevant to the workflow. The authority should be written before testing."
        },
        {
          "question": "Should sunk cost affect the decision?",
          "answer": "No. The decision should use expected future value, remaining risk, and the cost of the next credible evidence step—not money already spent."
        },
        {
          "question": "Is this legal or security advice?",
          "answer": "No. This provider-authored implementation guide should be adapted with the organisation’s legal, privacy, security, risk, finance, and sector specialists."
        }
      ]
    },
    "zhHant": {
      "name": "企業應在甚麼情況下停止 AI 項目？",
      "summary": "在試點開始前定義停止條件。當強制控制失敗、商業理據不再成立，或沒有具名負責人願意營運成果時，項目就應停止；不應只憑示範是否吸引作決定。",
      "decisionSignals": [
        {
          "name": "是決策，不是挫敗",
          "meaning": "保護資本及營運風險"
        },
        {
          "name": "強制關口",
          "meaning": "私隱、權限、安全及問責"
        },
        {
          "name": "留下證據",
          "meaning": "原因、負責人、處置及覆核日期"
        }
      ],
      "sections": [
        {
          "heading": "需要立即停止的條件",
          "introduction": "模型表現或簡報效果不能抵銷以下失敗。",
          "checks": [
            {
              "name": "權限失控",
              "requirement": "系統可在未批准的身份、租戶或權限邊界外存取受限資料或執行高影響動作。"
            },
            {
              "name": "商業理據失效",
              "requirement": "經核實的流程量、可實現收益、採用路徑或完整營運成本已不支持投資。"
            },
            {
              "name": "沒有問責負責人",
              "requirement": "交付團隊離場後，沒有具名業務負責人承擔流程、例外、剩餘風險及營運預算。"
            }
          ]
        },
        {
          "heading": "應暫緩並重新設計的訊號",
          "introduction": "目標可能仍然正確，但目前方案尚未產生可信證據時，應先暫緩。",
          "checks": [
            {
              "name": "數據或評估失敗",
              "requirement": "缺少代表性數據、標準答案不可靠，或測試集沒有涵蓋常見、邊界及高影響案例。"
            },
            {
              "name": "流程不匹配",
              "requirement": "操作人員繞過系統、人工覆核成本抵銷收益，或方案不符合真實決策節奏。"
            },
            {
              "name": "可靠性或成本越界",
              "requirement": "延遲、故障恢復、可用性或每次完整流程成本持續超出約定範圍。"
            }
          ]
        }
      ],
      "procedure": [
        {
          "position": 1,
          "name": "預先固定關口",
          "instruction": "建設前列明強制控制、量度門檻、證據負責人，以及誰可停止或重啟項目。"
        },
        {
          "position": 2,
          "name": "收集決策證據",
          "instruction": "把測試版本、輸入、結果、例外、人工介入、營運成本及使用情況放在同一可覆核記錄。"
        },
        {
          "position": 3,
          "name": "進行關口覆核",
          "instruction": "逐項標記通過、失敗、未知或未測試；平均分不可掩蓋強制關口失敗。"
        },
        {
          "position": 4,
          "name": "記錄處置決定",
          "instruction": "選擇擴大、重設、暫緩或停止，並記錄原因、剩餘責任、資產交接及下次覆核日期。"
        }
      ],
      "frequentlyAskedQuestions": [
        {
          "question": "停止 AI 試點是否代表失敗？",
          "answer": "不一定。若停止能避免不安全上線或繼續投資於已失效的商業理據，就是一個有效的受控結果。"
        },
        {
          "question": "誰應擁有停止權？",
          "answer": "至少包括業務負責人，以及與該流程相關的私隱、安全、保安或受監管風險負責人；權限應在測試前寫明。"
        },
        {
          "question": "已投入成本應否影響決定？",
          "answer": "不應。決定應基於未來預期價值、剩餘風險及取得下一項可信證據的成本，而非已經花掉的金額。"
        },
        {
          "question": "這是法律或安全意見嗎？",
          "answer": "不是。這是由服務商撰寫的實施指南，應由機構的法律、私隱、安全、風險、財務及行業專家調整。"
        }
      ]
    },
    "zhHans": {
      "name": "企业应该在什么情况下停止 AI 项目？",
      "summary": "在试点开始前定义停止条件。当强制控制失败、商业依据不再成立，或没有明确负责人愿意运营成果时，项目就应该停止；不能只看演示是否精彩。",
      "decisionSignals": [
        {
          "name": "是决策，不是挫败",
          "meaning": "保护资本和运营风险"
        },
        {
          "name": "强制关口",
          "meaning": "隐私、权限、安全和问责"
        },
        {
          "name": "保留证据",
          "meaning": "原因、负责人、处置和复核日期"
        }
      ],
      "sections": [
        {
          "heading": "需要立即停止的条件",
          "introduction": "模型表现或演示效果不能抵消以下失败。",
          "checks": [
            {
              "name": "权限失控",
              "requirement": "系统可以在未经批准的身份、租户或权限边界外访问受限数据或执行高影响动作。"
            },
            {
              "name": "商业依据失效",
              "requirement": "经过核实的流程量、可实现收益、采用路径或完整运营成本已经不支持投资。"
            },
            {
              "name": "没有问责负责人",
              "requirement": "交付团队离场后，没有明确业务负责人承担流程、异常、剩余风险和运营预算。"
            }
          ]
        },
        {
          "heading": "应该暂停并重新设计的信号",
          "introduction": "目标可能仍然正确，但当前方案还没有产生可靠证据时，应该先暂停。",
          "checks": [
            {
              "name": "数据或评估失败",
              "requirement": "缺少代表性数据、标准答案不可靠，或测试集没有覆盖常见、边界和高影响案例。"
            },
            {
              "name": "流程不匹配",
              "requirement": "操作人员绕过系统、人工复核成本抵消收益，或方案不符合真实决策节奏。"
            },
            {
              "name": "可靠性或成本越界",
              "requirement": "延迟、故障恢复、可用性或每次完整流程成本持续超出约定范围。"
            }
          ]
        }
      ],
      "procedure": [
        {
          "position": 1,
          "name": "预先固定关口",
          "instruction": "建设前说明强制控制、测量门槛、证据负责人，以及谁可以停止或重启项目。"
        },
        {
          "position": 2,
          "name": "收集决策证据",
          "instruction": "把测试版本、输入、结果、异常、人工干预、运营成本和使用情况放在同一个可复核记录中。"
        },
        {
          "position": 3,
          "name": "进行关口复核",
          "instruction": "逐项标记通过、失败、未知或未测试；平均分不能掩盖强制关口失败。"
        },
        {
          "position": 4,
          "name": "记录处置决定",
          "instruction": "选择扩大、重设、暂停或停止，并记录原因、剩余责任、资产交接和下次复核日期。"
        }
      ],
      "frequentlyAskedQuestions": [
        {
          "question": "停止 AI 试点是否代表失败？",
          "answer": "不一定。如果停止能够避免不安全上线或继续投资于已经失效的商业依据，就是一个有效的受控结果。"
        },
        {
          "question": "谁应该拥有停止权？",
          "answer": "至少包括业务负责人，以及与该流程相关的隐私、安全或受监管风险负责人；权限应该在测试前写明。"
        },
        {
          "question": "已经投入的成本应该影响决定吗？",
          "answer": "不应该。决定应该基于未来预期价值、剩余风险和取得下一项可靠证据的成本，而不是已经花掉的金额。"
        },
        {
          "question": "这是法律或安全意见吗？",
          "answer": "不是。这是由服务商编写的实施指南，应该由组织的法律、隐私、安全、风险、财务和行业专家调整。"
        }
      ]
    }
  },
  "limitations": [
    "This asset is provider-authored and is not independent validation or a claim of client results.",
    "It is an implementation aid, not legal, privacy, security, financial, procurement, or regulatory advice.",
    "Thresholds and controls must be adapted to the organisation, workflow, jurisdiction, and risk owners."
  ]
}