What it means
An API (Application Programming Interface) is a defined way for one software system to request data or services from another. In AI projects, APIs are how applications send requests to AI models and how AI systems read from or write to existing business systems.
Real-world example
A municipal appointment booking system offers an API. An AI assistant on the city's website uses it to check available slots for passport applications and, after the citizen confirms, to book one. The assistant never accesses the booking database directly; it can only do what the API allows.
Why it matters for AI adoption
APIs define what an AI system can and cannot do with other systems, which makes them central to both functionality and security. Many AI projects stall because the needed data or functions have no usable API, or because existing APIs were never designed for automated access at scale.
Questions to ask
- Which systems need to connect, and do they offer suitable APIs?
- What exactly can the AI system read or change through each API?