是不是给个坐标，AI 就出结果？和通用 AI 有什么不同？Do you just give AI a coordinate and get an answer? How is this different from general AI?
完全不同。通用 AI 给坐标也会出结果，但没有过程、没有依据。StrikOre 的工作流按地质专家的思考方式构建：先整合资料、判断成矿构造与类型，再识别优先潜力区、提出验证方向。全过程是白盒，有完整证据链，可以分模块由专家修改，最终由专家论证。Completely. A general AI will return something for a coordinate, but with no process and no basis. StrikOre's workflow is built the way a geologist reasons: integrate the data, determine structural setting and deposit type, then rank prospective areas and propose tests. It's a white box with a full evidence chain, editable module by module, and experts make the final argument.
AI 会不会漏判或误判？Can the AI miss things or get them wrong?
会。AI 和专家都会受资料质量、方法适用性与认知边界影响。AI 的价值在于扩大可比较的信息范围、提高资料处理效率、保留可追溯的证据，减少单一个人视角的偏差，而不是天然比专家更准。Yes. Both AI and experts are limited by data quality, method suitability and what is known. AI's value is a wider body of comparable evidence, faster data handling, a traceable record and less single-viewpoint bias — not being inherently more accurate than experts.
如果原始资料有误，AI 也会跟着错。所以系统设有数据准确度识别环节，对存疑数据先标注、加约束；专家也会到现场核实资料与实际是否一致。If the source data is wrong, the AI will be too. That's why questionable data is flagged and constrained first, and experts check on site that the records match the ground.
遥感适用于所有矿种吗？Does remote sensing work for every deposit type?
不是。遥感是前期低成本、易获取的一类证据，但只对部分矿种和矿床类型有效；覆盖厚、植被密的地区信号会失真。卫星数据最多反映地表约 20–30 厘米的情况，不能作为排除潜力区的唯一依据，必须与地质、物探、化探、钻探和现场验证综合使用。No. Remote sensing is cheap, accessible early evidence, but only for some commodities and deposit styles; thick cover and dense vegetation distort the signal. Satellite data reflects at most the top 20–30 cm, can never rule an area out on its own, and has to be combined with geology, geophysics, geochemistry, drilling and field checks.
你们用的是哪个模型？是不是数据越多越准？Which model do you use? Is more data always better?
当前按任务调用能力合适的第三方通用或专项模型，不绑定单一供应商。我们自研的是智能体工作流、项目资料组织、任务编排、证据交叉验证和决策流程。We call suitable third-party general or specialist models per task, with no single-vendor lock-in. What we build ourselves is the agent workflow, project data organisation, orchestration, evidence cross-checking and decision process.
数据不是越多越准。世界上没有两个相同的矿，堆入无关数据多是负样本，反而增加误判。关键是找到与这个矿区真正相关的资料。More data isn't automatically better. No two deposits are alike, and irrelevant data mostly adds noise and false leads. What matters is finding the data truly relevant to this district.
做分析需要我们提供什么？资料不全能做吗？What do we need to provide? Can you work with incomplete data?
需要原始数据而不是成果图：测量数据（坑道、钻孔点位、测斜、测深）、样品化验数据、地质图、以往报告、物化探数据和踏勘记录，这样才能重新梳理、重新成图。We need raw data rather than finished maps: surveys (workings, collars, downhole surveys), assays, geological maps, past reports, geophysical and geochemical data and field records — so everything can be re-worked and re-mapped.
资料少也可以开始。先用低成本的遥感初筛和地表人工采样判断成矿可能性；如果可能性不大，我们会直接建议换区、及时止损。You can start with little. Low-cost remote-sensing screening and hand sampling come first; if the potential looks weak, we'll say so and recommend moving on.
资料怎么保密？国内敏感数据怎么处理？How is our data protected? What about sensitive data in China?
签署保密协议后接收资料。每个矿区有独立的资料库与工作流，未经人工监督批准，项目之间数据不互通；客户资料只用于该项目，不作为公开案例。Data is received under NDA. Each district has its own data store and workflow, and nothing crosses between projects without supervised approval. Client data serves that project only and is never a public case study.
涉及敏感地质、测绘或矿山数据时，先按数据类型和敏感等级确定处理方式，采用数据隔离、本地化或离线处理方案，在授权范围内工作，不把受限数据上传至不合规的外部服务。For sensitive geological, survey or mine data, handling is decided by data type and sensitivity: isolated, localised or offline processing within the authorisation given, and restricted data never goes to non-compliant external services.
能直接给出资源量、储量或项目价值吗？Can you give resources, reserves or project value directly?
不能。早期成果是"哪些区域值得优先验证、依据是什么、风险在哪里"的潜力区研判，不代表已经确认成矿。资源量、储量、项目价值和采选经济性需要钻探、化验，并由地质、资源储量、采矿、选矿、经济等专业人员共同承担和签署。No. Early results tell you which areas deserve testing first, why, and what the risks are — they don't confirm mineralisation. Resources, reserves, value and mining/processing economics require drilling and assays, and must be signed off by qualified geology, resource, mining, processing and economics professionals.
项目已经在打钻、到中后期了，还有用吗？We're already drilling. Is this still useful?
有。可以把地质、遥感、物化探、钻孔和未见矿信息重新组织，做独立的桌面复核，再由地质专家把结果与现有报告逐项对比，找出资料矛盾、证据缺口和下一步验证方向，也可以在已知矿脉之外寻找新的解释方向。这属于技术复核和决策支持，不替代正式的资源储量评价。Yes. Geology, remote sensing, geophysics, geochemistry, drilling and barren holes can be re-organised into an independent desktop review, which experts compare item by item with your existing reports to find contradictions, gaps and next tests — including new interpretations beyond known veins. It is technical review and decision support, not a formal resource assessment.
现场取样谁来做？交付报告后你们还参与吗？Who does the field sampling? Do you stay involved after delivery?
第一阶段建议用当地人力做地表人工采样，我们给出在哪采、怎么采的具体要求；进一步验证由合作方地质专家或我们安排的专家到现场。槽探、钻探等工程由专业勘探队实施。Early on, local crews can take surface samples by hand following our specifications on where and how. Further checks are done by your geologists or experts we send; trenching and drilling are carried out by professional contractors.
会持续参与。采样、化验、物探、钻探的新结果回流后，我们更新证据、调整判断和验证优先级。持续服务的范围在项目协议中约定。Yes. As new sampling, assay, geophysical and drilling results come back, we update the evidence, the interpretation and the priorities. The scope of ongoing work is set in the project agreement.
一个项目要做多久？How long does a project take?
取决于资料完整度、目标区规模、项目阶段和现场采样检测条件，前人资料的矛盾点越多，周期也越长。我们会在看过资料清单后给出具体计划，不对外使用固定天数。It depends on data completeness, area size, project stage and field sampling conditions — the more conflicts in prior work, the longer it takes. We'll give a specific plan after reviewing your data list rather than quote a fixed number of days.
Try StrikOre 什么时候能用？When can we use Try StrikOre?
StrikOre 首版计划于 2026 年 12 月开放，以遥感为首个重点方向，具体时间以届时正式信息为准。想第一批试用，发邮件到 ken@strikore.com。The first release of StrikOre is planned for December 2026, leading with remote sensing; timing is subject to official announcement. To be among the first users, email ken@strikore.com.