AI Usage & Responsible AI

Blabbe is an AI shopping assistant designed specifically for Shopify stores. It is not a generic chatbot. It is a store-aware system built to guide shoppers through product discovery, answer pre-sale questions, and support better buying decisions.

This page explains how Blabbe uses AI, the boundaries of the system, and the principles guiding its design.

1. What Blabbe Is Designed To Do

Blabbe is built to assist with product discovery, answering questions about products and pricing, explaining store policies, and helping customers move from intent to purchase.

The assistant is designed for pre-sale guidance and decision support.

2. What Makes Blabbe Different

Blabbe combines three core inputs: Store Data, Shopper Context, and Structured Processing. This allows the assistant to understand intent, refine answers, and provide context-aware recommendations.

3. How AI Is Used

When a shopper sends a message, AI is used as a generation layer, not as an autonomous system. All outputs are constrained by available store data and system-defined logic.

1. The message is analyzed for intent
2. Relevant store data is retrieved
3. Context from the conversation is applied
4. AI generates a response
5. The response is delivered in real time

4. AI Boundaries and Limitations

Blabbe is intentionally limited in scope. The assistant does not access external systems beyond the store context and does not guarantee the accuracy of responses.

5. Human Oversight and Merchant Control

Blabbe is a tool — not a replacement for human judgment. Merchants remain responsible for store policies and final customer decisions.

6. Data Usage in AI Processing

AI processing is limited to the shopper’s message and relevant store data. Blabbe does not build long-term behavioral profiles or use conversations to train global models.

7. Consistency and Variability

AI-generated responses are dynamic, context-dependent, and probabilistic. Phrasing and recommendations may change based on the available information at the time of the request.

8. Safe Interaction Design

Blabbe includes safeguards like structured intent handling and validation of parameters to prevent unsupported actions and unsafe operations.

9. Handling Missing or Incomplete Data

When store data is unavailable, the assistant adapts its response and avoids making unsupported claims, failing safely and gracefully.

10. Continuous Improvement

Blabbe continuously improves through system-level enhancements and better retrieval logic, focusing on accuracy, relevance, and user experience.

11. Responsible AI Principles

12. Positioning of the Assistant

Blabbe should be understood as a product discovery guide and a conversational interface for store data. It is not a decision-making authority or a fully autonomous system.