构建 Jarvis Pro：先路由，后回答

Introduction

简介

The first Jarvis Pro prototype could produce answers that sounded right.

第一个Jarvis Pro原型能够生成听起来正确的答案。

That was the problem.

这就是问题所在。

One early answer looked polished: it named the merchant, summarized the week, and recommended pushing promotions before the next review. It was also wrong. The merchant’s order volume was down, but the sharper issue was operational: more outlets were paused and fulfilment had slipped. Sending more demand into that setup would have made the merchant look worse.

早期的一个回答看起来很完善：它指出了商家名称，总结了上周的情况，并建议在下次复盘前加大促销力度。但它也是错的。该商家的订单量下降了，但更尖锐的问题在于运营层面：更多门店被暂停营业，履约能力也出现了下滑。在这种情况下引入更多需求只会让商家的处境更糟。

That failure changed how we judged the system. Fluent was not enough.

那次失败改变了我们评估系统的方式。仅仅表达流畅是不够的。

Jarvis Pro is the AI assistant we built for Grab account managers. Its job is to help them turn account data into better merchant conversations: what changed, why it changed, and what to do next. They rarely ask clean dashboard questions. They ask: “I am meeting this merchant tomorrow. What should I tell them?” or “Which accounts in my portfolio need attention this week?”

Jarvis Pro 是我们为 Grab 客户经理构建的 AI 助手。它的工作是帮助他们将账户数据转化为更高效的商户沟通：发生了什么变化，为什么变化，以及接下来该怎么做。他们很少提出明确的仪表盘问题。他们会问：“我明天要见这位商户。我应该告诉他们什么？”或者“我的客户组合中哪些账户本周需要关注？”

Those questions hide decisions: scope, access, business diagnosis, and metric definition. If the system gets those wrong, confidence becomes a liability.

这些问题背后隐藏着决策：范围、权限、业务诊断和指标定义。如果系统在这些方面出错，信心就会变成负担。

So the core design became: route first, answer later.

因此，核心设计变成了：先路由，后回答。

In an internal offline evaluation (not a measure of production performance or business impact), routing matched the expected safe route for 99.4% of 351 realistic prompts drawn from labelled eval sets from the first half of 2026. In a focused portfolio and brand answer-quality suite, the average score moved from 78.5 to 91.0. These figures come from offline launch-readiness evaluation only; they are not business-impact proof.

在一项内部离线评估（并非衡量生产性能或业务影响的指标）中，对于从2026年上半年标注评估集中提取的351个实际提示词，路由在99.4%的情况下匹配了预期的安全路由。在聚...