hoangsonww/RAG-LangChain-AI-System

Feature: Proactive Portfolio Watchlists and Scheduled Insight Briefs

Open

#22 opened on Mar 8, 2026

View on GitHub
 (0 comments) (0 reactions) (1 assignee)Jupyter Notebook (13 forks)auto 404
bugdocumentationenhancementgood first issuehelp wantedquestion

Repository metrics

Stars
 (46 stars)
PR merge metrics
 (PR metrics pending)

Description

Summary

Add proactive intelligence workflows: users define watchlists (companies, sectors, themes), and the system generates scheduled briefs with changes, risks, and opportunities using both retrieved documents and structured backend data.

Why this matters

Current chat is reactive. Portfolio teams need recurring intelligence without manually asking the same questions every day/week.

Scope

  • Watchlist configuration model (entities + alert rules + cadence).
  • Scheduled brief generation jobs (daily/weekly).
  • Delta-aware summaries (what changed since last brief).
  • Delivery endpoints (in-app feed; optional email/webhook stub).
  • Brief traceability: cited sources and tool calls included.

Non-goals

  • Full notification provider integration matrix in v1.
  • External market data vendor integrations in this issue.

Proposed implementation

  1. Add watchlist CRUD APIs and frontend settings UI.
  2. Build scheduler that runs templated analyses per watchlist.
  3. Implement “change since last run” comparison for concise deltas.
  4. Store generated briefs as queryable artifacts tied to workspace/user.
  5. Add brief viewer UI with source cards and confidence metadata.

Acceptance criteria

  • Users can create, edit, and pause watchlists.
  • Scheduled jobs generate briefs at configured cadence.
  • Briefs include explicit sections: key changes, opportunities, risks, follow-up questions.
  • Every brief has linked evidence sources and tool-trace metadata.
  • Missed/failed schedules are retried and surfaced in job status.
  • End-to-end tests validate generation + retrieval provenance for briefs.

Labels

enhancement, product-feature, backend, frontend

Contributor guide