Guardrails
The limits placed on what an AI may say or do — refusing certain requests, staying on topic, never inventing a price. Essential on anything customer-facing.
Related terms
Model Drift
When a model gets worse over time because the world moved on and it did not. A pricing or demand model trained two years ago quietly stops matching how customers behave now.
Deep Learning
Machine learning using neural networks with many layers. The depth is what lets a system learn complicated things — recognising a face, translating a sentence — without anyone specifying how.
E-E-A-T
Google’s shorthand for Experience, Expertise, Authoritativeness and Trust — the qualities its raters look for. Not a score you can see, but a decent description of what makes content worth surfacing.
AI Visibility
How often, and how accurately, AI assistants mention your business when someone asks about what you do. Unlike search rankings there is no official scoreboard, so it is measured by asking the assistants and by watching which crawlers visit.
Agentic Browsing
An AI navigating a website on somebody’s behalf: finding the price, checking availability, sometimes buying. Sites that rely on hover menus and unlabelled buttons are hard for an agent to use.
Prompt Engineering
The craft of writing instructions that get reliable results from an AI — being specific, giving examples, saying what to avoid. Less mysterious than it sounds, and mostly the same skill as briefing a new colleague well.