ARIA

The description states that ARIA is a multi-agent autonomous system designed to provide Nigerian SME merchants with intelligent business insights by analyzing unstructured data from WhatsApp messages and bank alerts. The system operates continuously in the background, generating daily briefings with three prioritized actions. It uses LangGraph for orchestration, GPT-5.6 for reasoning, Codex for dynamic script generation, and FastAPI for backend services.

The project is self-reported as a proof-of-concept built during an OpenAI hackathon. No revenue, customers or traction data are provided. The description indicates the team has not yet launched commercially, but they have built a working prototype that demonstrates core functionality including autonomous analysis code generation and decision-layer filtering of insights.

The description states ARIA is:

• A "continuously running multi-agent system" (not a chatbot or prompt loop)
• Designed to work in the background monitoring merchant data
• A "proactive daily briefing" with exactly three prioritized actions
• Built using LangGraph, GPT-5.6, Codex, FastAPI, PostgreSQL, Python

The system has four agents:

1. Ingestion Agent - parses SMS bank alerts and WhatsApp messages into structured events
2. Codex Analysis Agent - generates bespoke Python analysis scripts at runtime
3. GPT-5.6 Intelligence Agent - reasons across accumulated signals to identify implications
4. Decision Layer - filters insights for actionability, urgency, impact and relevance

The description states this is not a chatbot but an autonomous agent that performs genuine business work without user prompts after initial data upload.

The description states ARIA positions itself as:

• An "autonomous revenue intelligence agent"
• Working while Nigerian merchants sleep
• Monitoring signals, writing its own analysis code, and delivering 3 clear actions daily
• Filling a gap for 41 million Nigerian SMEs with no analytics team, CRM or intelligence layer

The claim evolution appears to be:

1. Start with the problem: Nigerian merchants make decisions on gut instinct without data
2. Present solution: ARIA analyzes existing WhatsApp and bank alert data intelligently
3. Demonstrate capability: Shows autonomous operation with dynamic script generation
4. Highlight unique value: Decision layer filters insights to only what merchant needs to act on

The positioning is described as targeting "Nigerian SMEs" specifically, with the Nigerian context being the "proof of concept" and architecture being market-agnostic.

The description states:

• Primary target: Nigerian SME merchants
• Specific use case: 41 million Nigerian SMEs with no analytics team, CRM or intelligence layer
• Business decision context: Who to follow up with, what to restock, which customer is drifting
• Data sources: WhatsApp messages and bank alerts

The description does not state:

• Specific merchant size or revenue thresholds
• Industry verticals beyond general SMEs
• Geographic expansion plans beyond Nigeria
• Customer segmentation beyond "Nigerian SME"

Not evidenced. The description states no commercial data, revenue or pricing information is available.

The description states:

• Built with LangGraph for multi-agent orchestration
• Uses GPT-5.6 for cross-signal reasoning
• Codex for live analysis script generation
• FastAPI for backend
• PostgreSQL for event storage
• Python for execution
• All core functions built through Codex inside ChatGPT
• The Codex Analysis Agent itself was architected and built entirely through Codex sessions

The description states:

• Four agents run in a LangGraph pipeline
• Ingestion agent parses raw SMS and WhatsApp into structured events
• Codex Analysis Agent generates and executes bespoke Python scripts
• GPT-5.6 Intelligence Agent reasons across signals to identify implications
• Decision Layer filters insights before delivery

Not evidenced. The description states:

• This is a hackathon project submitted to OpenAI 2026 hackathon
• No revenue, customer or traction data available beyond the authors' account
• The team has not yet launched commercially
• They have built a working prototype that demonstrates core functionality
• The moment that made the build feel real was watching it process 30 days of raw data and produce a complete business briefing without user prompts

Not evidenced. The description does not state:

• Direct competitors
• Market size or competitive landscape
• Existing solutions in this space
• Differentiation from similar tools

The description states:

• The hardest problem was the decision layer - generating intelligence is tractable but knowing what NOT to surface is difficult
• The second challenge was execution safety for Codex Analysis Agent - needed sandboxing, timeout enforcement and output validation
• No commercial traction or revenue data provided
• The system is described as a hackathon prototype with no indication of production readiness
• The team size is 3 people, which may limit development capacity

1. What specific business decisions do merchants make that ARIA helps with?
2. How does the decision layer actually work in practice? What are the filtering criteria?
3. What are the actual technical challenges around sandboxing and safety of Codex-generated code?
4. Has any merchant actually used this system beyond the prototype phase?
5. What is the current state of the WhatsApp Business API integration?
6. How do you plan to scale from a 3-person team to serving thousands of merchants?
7. What are the actual costs of running this system at scale?