[Supplier] Phase 6: AI Price Researcher Service & Lean Modal (Least Priority) #200

Open
opened 2026-10-03 23:22:23 +00:00 by yuris · 0 comments
Owner

Overview

Implement an on-demand market price research service that queries an LLM with web search capabilities behind the scenes to fetch official retail benchmarks (HET/SPBU) and verified marketplace pricing, presenting a clean, uncrowded recommendation modal to assist pricing decisions.

Note: Least priority phase. Scheduled after Phases 1–5 are shipped and stabilized.

Principles Applied

  • principle-laziness-protocol: Lean popover instead of complex 10-widget analytics dashboards.
  • principle-boundary-discipline: Strict validation and anti-counterfeit filtering of search outputs.

Scope & Tasks

  • Implement backend service in apps/management/lib/price-research/service.ts querying LLM/web-search.
  • Input context: item name, incoming cost, workshop location (Bandung, Jawa Barat).
  • Anti-counterfeit filter: discard unverified low-cost marketplace listings to prevent contaminated benchmarks.
  • Output schema: online mall price, official HET/SPBU price, recommended garage price, single-sentence rationale, and 2-3 verified citations.
  • Lean UI modal displaying:
    1. Market Anchors (Online Mall vs SPBU HET).
    2. Recommended Store Price (with profit nominal & margin %).
    3. One-sentence commercial rationale.
    4. 2–3 verified citations.
    5. [ Terapkan ] button filling the row's sell price input.

Acceptance Criteria

  • Clicking [ ✦ Riset AI ] loads benchmarks within 3–5 seconds.
  • UI is clean, focused, and free of redundant charts, gauges, and seller card spam.
  • Clicking Apply populates the recommended price directly into the catalog row.
  • pnpm verify:management passes.
### Overview Implement an on-demand market price research service that queries an LLM with web search capabilities behind the scenes to fetch official retail benchmarks (HET/SPBU) and verified marketplace pricing, presenting a clean, uncrowded recommendation modal to assist pricing decisions. *Note: Least priority phase. Scheduled after Phases 1–5 are shipped and stabilized.* ### Principles Applied - `principle-laziness-protocol`: Lean popover instead of complex 10-widget analytics dashboards. - `principle-boundary-discipline`: Strict validation and anti-counterfeit filtering of search outputs. ### Scope & Tasks - Implement backend service in `apps/management/lib/price-research/service.ts` querying LLM/web-search. - Input context: item name, incoming cost, workshop location (Bandung, Jawa Barat). - Anti-counterfeit filter: discard unverified low-cost marketplace listings to prevent contaminated benchmarks. - Output schema: online mall price, official HET/SPBU price, recommended garage price, single-sentence rationale, and 2-3 verified citations. - Lean UI modal displaying: 1. Market Anchors (Online Mall vs SPBU HET). 2. Recommended Store Price (with profit nominal & margin %). 3. One-sentence commercial rationale. 4. 2–3 verified citations. 5. `[ Terapkan ]` button filling the row's sell price input. ### Acceptance Criteria - [ ] Clicking `[ ✦ Riset AI ]` loads benchmarks within 3–5 seconds. - [ ] UI is clean, focused, and free of redundant charts, gauges, and seller card spam. - [ ] Clicking Apply populates the recommended price directly into the catalog row. - [ ] `pnpm verify:management` passes.
yuris changed title from [Pemasok] Phase 6: AI Price Researcher Service & Lean Modal (Least Priority) to [Supplier] Phase 6: AI Price Researcher Service & Lean Modal (Least Priority) 2026-10-03 23:23:17 +00:00
Sign in to join this conversation.
No labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
yuris/garage#200
No description provided.