What is AIML?
AIML is a legacy language that’s been around a while now – basically a simple way to script rules for chatbots, or put another way, a way to tell a chatbot what to do when a user says something particular. It’s a pretty old technique that was popular back in the early 2000s with the likes of A.L.I.C.E. chatbot.
Note: AIML is its own thing, a specific scripting language, and isn’t the same as ‘AI/ML’ (artificial intelligence and machine learning) – even though they bear a certain resemblance.
How is AIML used?
When developers write chatbots, they usually write out a bunch of rules in AIML files. These rules are basically saying “if the user says this, then reply with that”. It’s all pretty straightforward and old-school – it relies on exact matches between what the user says and a pre-written pattern. It’s not really “learning” from conversations like newer chatbots do.
How Onextel’s chatbots compare
Our current chatbot and conversational solutions are built on some pretty modern technology – we’re talking NLP and AI-based understanding that can handle natural language as it’s spoken, rather than relying on a bunch of pre-written scripts like AIML does. That means they can cope with all sorts of different ways of saying things, rather than just looking for an exact match.
Why AIML still matters
To be honest, AIML is mostly of historical interest at this point – it shows just how much chatbots have evolved since those early days. We’ve moved from a time when chatbots had to be told exactly what to say in every situation, to now where we’ve got AI chatbots that can understand the context of a conversation and respond accordingly.