.png)
One superpower that always seemed interesting is the ability to read people’s minds. If you could only understand what people meant without further follow-up questions, then it would be a convenient way to have conversations. But what if that is somewhat achievable with the help of AI? Learning from past conversations allows it to look into patterns and then letting you know what the other person actually wants.
Let’s put it in the context of a commerce-related chatbot where you have an involvement with every step or prompt you to choose from a set of options before moving on to the next step. For example, you want to know where your parcel is. First, you need to go through the main menu then choose the sub-option for queries related to Logistics and then finally choose the Order Tracking option. This is necessary because chatbots rely on these to redirect you to the appropriate link or instructions. A smart chatbot could skip some steps if it can already identify the intent with just a single message.
Smart conversational AI has been trained to determine what a specific phrase means based on past conversations. It looks for the usual pattern found in a certain intent and is able to guess which intent a query falls under from that. For example, “where is my parcel”, “track my parcel”, “where is my order?”, these are commonly asked by users to know where their order is. If you train an AI with more of these phrases, then it can determine an intent with a similar context like “where is my package?” even if it’s the first time it has seen this phrase. Since we already know the intent from the message itself, we won’t need to ask users to click on menu options to redirect them, instead we can automatically do this for them
This same technology is being utilized by the Intent Classification API powered by aiah.ai. This service is accessible through this link: Intent Classification Service. You can also try this service for free as long as you sign up for an account. Once you’ve tested it out first, you’ll know if this is something that fits what you’re looking for. You can also find the documentation below which includes the scope of the trained AI model which is limited to commerce-related queries for now.
All you have to do is to enter your query in the “text” area of the request body and click “test endpoint”. From the example below, we can try a sample query like “where is my parcel?”. The results will show you the possible intent of the user based on this one query. It’s already ranked considering its relevance to the text itself.
If you’re a developer who wants to make use of this API and add the functionality to your existing app, you can use the snippet below to query from the service itself straight from the AI marketplace. As long as you have an account and connected project, you can use this as much as you want.
That’s it, that’s the whole snippet! If you tried the same query as above, the results should look like this:
You’re not limited to just using Python to be able to integrate it in your app, you can also choose from the available snippets for other programming languages found in the AI Marketplace Playground.
If you're looking to add AI to your app or business, go to aiah and discover and try services like Sentiment Analysis, Intent Route, Document Extraction, or any of our partner services.
