Logistics dev seeks architecture advice for internal AI assistant over company sales data
Dry_Tune_2432 · reddit · 2026-10-12
A developer at a logistics company is asking Reddit how to architect an internal AI assistant that lets employees query sales and ops data in natural language (monthly sales, comparisons, top customers) instead of digging through dashboards.
Key open questions:
- Direct DB access vs. a controlled API layer; is text-to-SQL appropriate; is RAG even needed for structured data?
- Per-employee data permissions, blocking unsafe or expensive queries, and preventing hallucinated financial numbers
- Model choice (OpenAI/Claude/Gemini/local) and handling sensitive data with external LLM APIs; connect to an existing BI system instead of production DB?
His draft pattern: employee chat → LLM → backend API → database/ERP → AI summary. He explicitly wants secure, maintainable, reliable design rather than AI for AI's sake — a good thread for anyone who has shipped similar text-to-SQL + permission-gateway setups.
More from coding & agent
- Ice Cubes dev uses Codex to clear open-source repo backlog of bumps and fixes — Dimillian · 2026-10-12
- SEO team uses Grok bot groups as web crawler, sometimes beating ScreamingFrog locally — gaganghotra_ · 2026-10-12
- EXL3-QLoRA lets you fine-tune 2-8bpw quantized models, beating BNB NF4 on loss — Unstable_Llama · 2026-10-12
- Linus Torvalds is using Claude Code on a hefty Linux kernel performance improvement — PointmanW · 2026-10-12
- Claude-driven Blender 3D modeling from images makes this dev bullish on inspection gear — jwt0625 · 2026-10-12
- Work2Skill Turns Human GUI Demos into Reusable Agent Skills, Beating No-Skill Baselines — nex-agi · 2026-10-12