--- title: "Building Jarvis Pro: Route first, answer later" url: https://daily.dev/posts/building-jarvis-pro-route-first-answer-later-rytlobkw5 source_url: https://engineering.grab.com/jarvis-pro-route-firsr-answer-later type: article source: "Grab Tech Blog" published: 2026-08-21T08:33:30.904Z updated: 2026-08-21T08:33:55.406Z tags: ["llm", "ai-agents", "rag"] reading_time: 7 upvotes: 0 comments: 0 language: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Building Jarvis Pro: Route first, answer later **[Grab Tech Blog](https://daily.dev/sources/grab)** · 7 min read · 0 upvotes · 0 comments ## Summary Grab's engineering team describes building Jarvis Pro, an internal AI assistant for account managers that routes user questions to a specific task type before generating an answer, rather than generating a fluent response directly. The design addresses a failure mode where an early prototype produced polished but operationally wrong recommendations by missing that a merchant's real problem was paused outlets and fulfilment issues, not demand. The system pipeline runs route classification, memory/access checks, context loading, and metric reconciliation before generation, deliberately trading a slower first token for safer answers. Offline evaluation on 351 routing prompts hit 99.4% expected-route match, and a 150-case portfolio/brand answer-quality subset improved from 78.5 to 91.0 on a 0-100 rubric; these are launch-readiness offline metrics, not proof of business impact. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### Why did Grab add a routing step before letting an LLM answer account manager questions in Jarvis Pro? An early prototype produced a fluent but wrong recommendation because it missed that a merchant's real issue was paused outlets and fulfilment failure, not demand, and pushing promotions would have made things worse. Routing first lets the system classify the task, load only relevant context, apply guardrails, and avoid jumping straight to generic recommendations like discounts. _See how daily.dev surfaces engineering writeups like this on building safer LLM agent pipelines._ ### What evaluation results did Grab report for the Jarvis Pro routing system before launch? Routing matched the expected safe route in 99.4% of 351 labelled offline evaluation prompts, and a 150-case portfolio and brand answer-quality subset improved from an average score of 78.5 to 91.0 on a 0-100 rubric covering template fit, metric use, diagnosis, and guardrail compliance. These are offline launch-readiness figures, not measures of commercial impact. _Compare LLM evaluation approaches like this one on daily.dev when benchmarking your own agent pipeline._ ## Similar posts on daily.dev - [Real-Time Customer Support Agent with Fallback Routing](https://daily.dev/posts/real-time-customer-support-agent-with-fallback-routing-kaqdajqmh) · DigitalOcean Community · 1 upvotes · 0 comments - [Building Biz Ask Anything: From Prototype to Product](https://daily.dev/posts/building-biz-ask-anything-from-prototype-to-product-9efe0mz1d) · Yelp Engineering · 0 upvotes · 0 comments - [Cut OpenClaw Bill by 50%](https://daily.dev/posts/cut-openclaw-bill-by-50--a1tpaby9m) · Daily Dose of Data Science \| Avi Chawla \| Substack · 2 upvotes · 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#ai-agents](https://daily.dev/tags/ai-agents), [#rag](https://daily.dev/tags/rag) [View this post on daily.dev](https://daily.dev/posts/building-jarvis-pro-route-first-answer-later-rytlobkw5)