gokilaharini / Custos_AI

What is Custos_AI?

Custos_AI is a Python-based Streamlit app powered by LangChain and Groq LLMs that lets telecom teams analyze customer churn risks, assess financial exposure like total monthly revenue at stake, and generate data-driven retention decisions via natural language queries. You ingest a CSV dataset into SQLite, fire up the web UI with `streamlit run run.py`, and ask things like "Analyze customer 7795-CFOCW" or "Make a decision for this churn-risk user"—it pulls profiles, crunches numbers, and even stores recommendations in a decision log. Built for custos ai scenarios where understanding customer churn and business costs directly impacts retention.

Why is it gaining traction?

It stands out with an autonomous agent that not only analyzes but acts—routing queries to customer risk profiles, finance summaries, or full decisions—without manual SQL tweaking. Developers dig the end-to-end workflow: quick data ingestion, intent detection for queries, and persistent memory for repeat business decisions, making it a solid LangChain agent demo for custos openai or custos api open ai integrations. Low setup barrier hooks prototyper who want agentic AI without building from scratch.

Who should use this?

Telecom data analysts evaluating churn models with real datasets, ML engineers prototyping agentic workflows for customer retention, or business intelligence devs needing fast financial risk assessments like average monthly charges across high-churn users. Ideal for teams exploring custos ventures aif or custos buenos aires-style cost optimization in telecom.

Verdict

With 31 stars and a 1.0% credibility score, it's an early-stage project—docs are basic, no tests visible—but a constructive starter for LangChain agent experiments in churn analysis. Try it if you're dipping into data-driven decision agents; skip for production without hardening. (187 words)