Amigo AI

Landing page

Inspiration

We were inspired by a simple problem: emerging markets generate enormous investment opportunities, but most of the intelligence is fragmented, multilingual, and difficult to analyze quickly. Professional analysts spend hours reading filings, monitoring news, validating macroeconomic trends, and manually connecting signals across different sources. Existing AI tools are heavily focused on U.S. markets and often generate hallucinated or ungrounded financial insights. We wanted to build an autonomous AI research analyst capable of discovering opportunities in underfollowed markets like Brazil and transforming scattered information into actionable investment intelligence. Amigo AI demonstrates how agentic AI systems can automate market research workflows while remaining grounded in real financial documents, reports, and structured analytics.

What it does

Amigo AI is an AI-powered market intelligence platform focused on emerging markets, starting with Brazil. The platform autonomously:

Discovers relevant market signals from news, filings, and reports
Classifies investment relevance and sentiment
Stores structured analytics in a scalable database
Retrieves grounded financial context using semantic search
Generates investment memos, opportunity rankings, and market summaries
Monetizes premium research through autonomous API payments Initial focus sectors:
Brazilian agriculture
Logistics
Energy
Infrastructure Target users:
Hedge funds
Investment analysts
PE/VC firms
Commodity traders
Macro research teams

How we built it

Discovery Layer

We used Nimble to autonomously search and crawl:

Brazilian market news
Financial filings
Economic reports
Sector-specific intelligence

The discovery agent extracts:

Companies
Sectors
Sentiment
Macroeconomic signals
Investment relevance

Analytics Layer

We integrated ClickHouse to store and analyze:

Financial metrics
Commodity prices
Sentiment scores
Market signals
Structured economic data

Semantic Memory Layer

Using Senso.ai, we implemented grounded retrieval over:

Earnings reports
Filings
PDFs
Policy documents
Historical news archives This ensures generated insights remain tied to real sources instead of hallucinated outputs.

Analysis Agent

Our analysis agent combines:

Live web signals
Structured analytics
Semantic retrieval to generate:
Investment theses
Bullish/bearish outlooks
Sector summaries
Opportunity rankings

Challenges we ran into

One of the biggest challenges was coordinating multiple AI agents and infrastructure components within a hackathon timeframe. We had to solve:

Multi-agent orchestration
Structured + unstructured data integration
Retrieval grounding
API interoperability
Scalable analytics pipelines Another major challenge was ensuring the AI generated credible financial insights instead of speculative or hallucinated outputs. We addressed this through grounded semantic retrieval using real filings and reports. Working with emerging-market data also introduced difficulties because many sources are fragmented, multilingual, and inconsistent in structure. Finally, balancing technical ambition with realistic MVP scope was critical. We intentionally focused on Brazil and a limited number of sectors to deliver a polished end-to-end demo. ---

Accomplishments that we're proud of

We are especially proud of building an end-to-end autonomous market intelligence platform within a hackathon environment while integrating multiple advanced AI and analytics technologies into a single workflow. Some of our biggest accomplishments include: Building a multi-agent AI system capable of autonomously discovering and analyzing emerging-market investment opportunities Creating grounded financial intelligence workflows using semantic retrieval instead of unreliable hallucinated outputs Successfully integrating structured analytics with unstructured market data from news, reports, and filings Implementing a scalable architecture using Nimble, ClickHouse, Senso.ai, Datadog, Lapdog, and x402 Demonstrating autonomous monetization through machine-payments and pay-per-analysis APIs Designing a focused MVP around Brazilian agriculture and logistics with real-world commercial relevance Creating a platform that combines AI research, observability, analytics, and monetization into one cohesive product Delivering institutional-style investment intelligence workflows that could scale across Latin America and other emerging markets We are also proud of balancing technical ambition with practical execution by narrowing the scope and delivering a polished, credible demo in a short amount of time.

What we learned

During development we learned how powerful agentic AI systems can become when combined with:

Retrieval-augmented generation
Structured analytics
Observability tooling
Autonomous workflows We also learned that financial AI products require far more than just LLM prompting. Reliable investment intelligence depends on:
Grounded data retrieval
Traceability
Data quality
Workflow orchestration
Scalable infrastructure The project also gave us hands-on experience integrating AI systems with monetization infrastructure and analytics pipelines.

What's next for Amigo AI

Future plans for Amigo AI include:

Multi-country expansion across Latin America
Portuguese-native ingestion and analysis
Autonomous investor alerts
Geopolitical monitoring
Predictive forecasting models
Portfolio intelligence features
Institutional integrations
Proprietary market scoring systems Our long-term vision is to create an autonomous AI research platform capable of delivering institutional-grade emerging-market intelligence globally.