We often talk about artificial intelligence (AI) in pretty broad terms, but as you know well, AI is kind of a misnomer that is really a compilation of a number of different technologies. TechRepublic spoke with Gupshup co-founder and CEO Beerud Sheth to discuss what types of machine learning, data analysis, and other predictive tools power an AI chatbot.
Sheth: AI is a collection of a lot of different techniques and such. One of the commonest things that's used in chatbots vs. natural language processing, which is used for understanding regular, unstructured queries. Let's say if a user wants to buy a ticket, they may say the same thing a dozen different ways. I wanna make a booking. I want to reserve. I want to fly to New York. Whatever it is. A natural language processing technique or algorithm would help you figure out what the user is saying.
SEE: Research: Companies lack skills to implement and support AI and machine learning (Tech Pro Research)
Another thing that people use often is for personalization. Based on their past track record, we can make predictions about what they are likely to want or need next. We've done that, for example, predicting the kind of news stories they prefer. You can use it for things like pattern recognition and image detection and so on. The user sends a photo and now the bot has to figure out what is it. Is there a cat in the photo or not in it. Or let's say if it's a receipt of a restaurant, you want to be able to read what's in there, and classify it correctly and so on.
And then getting into more advanced things. There's a lot of things, deep learning and so on. Those could come in. The benefit of AI in general is that the bot can engage with humans in a convenient way. Like I said, in terms of the user having to identify the receipt and the restaurant name and the amount, the bot can look it up automatically. He just sends a photo, and the bot can figure out what's in the photo, etc.
So, it makes it easy and convenient, and these technologies—as they evolve—they keep getting better. Of course, for voices, voice recognition is a big deal, another big AI technique, and you know these things are getting better every day.
- Understanding the differences between AI, machine learning, and deep learning (TechRepublic)
- Machine learning: The smart person's guide (TechRepublic)
- IBM Watson: Six lessons from an early adopter on how to do machine learning (TechRepublic)
- Your customers don't want to talk to you, they want to talk to your chatbot (TechRepublic)
- AI chatbots and voice assistants growing as tools for shopping, paying bills, banking (TechRepublic)
Dan Patterson has nothing to disclose. He does not hold investments in the technology companies he covers.
Dan is a Senior Writer for TechRepublic. He covers cybersecurity and the intersection of technology, politics and government.