采购方工作区
AI 辅助评估 — 分析自动完成，决定由您作出
开标记录完成后，每份投标都会按您已公布的标准进行核查。工作区展示各投标人说了什么、出处在哪里以及缺少什么。系统不作任何决定：评估人员评分，委员会决策，记录明确显示由谁作出了什么决定。
- 每项强制性要求都会对照已提交文件核查，每项发现均附页级引用。
- 评分开始前标出缺失、过期或无法读取的文件，以便及早发出澄清请求，而非等到截止日期之后。
- 根据价格、合规情况及您定义的标准，并排比较各投标人，无需导出电子表格。
- 评估人员自行输入分数和书面理由；系统绝不会自行给出最终分数。
- 共识评分将每位评估人员的原始记录、协调后的结果及其理由保存在同一记录中。
- 完整评估记录，包括发现、引用、分数、理由及批准人，始终与招标关联，以供复核或审计。
Ranked submissions
Closed 14 October · 4 bidsSupplier A
Meets all mandatory criteria; strongest technical response
Supplier B
Compliant; fewer comparable references
Supplier C
Pricing schedule incomplete — clarification required
Supplier D
Excluded: mandatory certificate expired
Compliance checked before scoring
Requirements are read straight from your tender documents, so the checklist matches what you actually published rather than a generic template.
18 requirements extracted
Export CSVValid tax clearance certificate
Page 7, clause 5.1 — mandatory
ISO 9001 quality certification
Page 8, clause 5.4 — mandatory
Three references, similar scope
Page 9, clause 6.2 — scored
Local content declaration
Page 11, annexure C — mandatory
Registered supplier database entry
Page 6, clause 4.3 — mandatory
Gap detected
No local content declaration found in your documents. Required before submission.
Where the line sits
The analysis does the reading: hundreds of pages per submission, checked against the published criteria and cited back to the source. People do the judging. No award is recommended, ranked or approved automatically, and no score is written without a named evaluator behind it. That separation is what keeps an evaluation defensible.