Navigating AI support tools is easier when you understand the vocabulary. Here are the 30 most important terms, explained clearly for non-technical ecommerce founders.
The world of AI support tools comes with a lot of jargon. Understanding the terminology helps you evaluate tools more accurately, ask better questions of vendors, and make better decisions about your support infrastructure.
Core AI concepts
- LLM (Large Language Model): The AI that reads and generates text. GPT-4 and Claude are LLMs. The quality of the LLM is the primary determinant of reply quality.
- RAG (Retrieval-Augmented Generation): The technique that combines an AI's language ability with your specific knowledge base. This is how AI answers questions about your specific store instead of giving generic answers.
- Token: The basic unit of AI processing. Roughly 3/4 of a word. Used in pricing models to measure usage.
- Confidence threshold: The minimum certainty an AI needs to send a reply automatically without human review. Set too low = errors. Set too high = too many escalations.
- Escalation: When the AI determines it cannot handle a ticket confidently and passes it to a human.
Support operations terms
- WISMO: Where Is My Order. The most common Shopify support ticket category.
- FRT: First Response Time. Time between ticket receipt and first reply. The most important support metric.
- CSAT: Customer Satisfaction Score. Typically rated 1–5 after a support interaction.
- Resolution rate: Percentage of tickets fully resolved in a single interaction.
- Ticket deflection: Preventing a support ticket from being created in the first place (via FAQ, proactive notifications, etc.).
- Automation rate: Percentage of tickets handled without human intervention.
