There’s a Bot for That: Chatbot & Artificial Intelligence Fundamentals

Engineering, AI/Bots

Tweet about this on TwitterShare on FacebookShare on LinkedInShare on Google+

Blog-Chatbots-Machine-Learning-incontent

Remember chatting with SmarterChild on AIM? What about the ahead-of-its time (but kind of creepy) Spielberg movie A.I. Artificial Intelligence? It wasn’t so long ago that artificial intelligence seemed a little too futuristic. Fast-forward to 2016, and AI-powered bots have become the next piece of emerging technology. Are they here to stay?

Some bot basics

“Bots” typically refers to a software that automates tasks. The most prolific bots—Google Assistant, Apple Siri, Amazon Alexa and Facebook M—are among major players in the domain. However, this does not mean the market has remotely approached its potential. To understand the end-user benefit of bots, consider Facebook Messenger.

Facebook Messenger acts as a type of bot portal with over 30,000 active bots as of October, 2016, most of which set the standard for current bot functionality. For example, when you input a query into Messenger, such as, “When is the next Caltrain from San Francisco to San Jose,” this hypothetical Caltrain bot could reply with a text answer, a graphic of the schedule and ticket purchase options. The consumer and business develop a virtually instant relationship that satisfies both parties—the basics of CRM in practice.

How exactly do bots, bot?

Developers strive to make it difficult for users to distinguish bot programs from a person—known as passing the Turing test—and to do that the bot needs to perceive its environment and make decisions. Enter Artificial Intelligence (AI).

In its simplest form, artificial intelligence simulates human intelligence by responding to questions based on encoded sets of data. For example, if you ask a bot the location of the closest market, it references its answers and delivers the most accurate response based on your location. But this level of interaction doesn’t intimate the AI-driven potential of conversational commerce.

Artificial intelligence is often enabled by way of machine learning, another word you’re likely hearing all over the place. Machine learning provides the framework through which a software program can test various data sets and responses against a specific outcome. Instead of manually programming questions and responses into a script for a bot to reference when interacting with a customer, the developer devises algorithms that “teach” a bot how to respond instead of referring to a boilerplate response.

Where does machine learning come in?

Machine learning will allow for personality-based bots, or bots that reflect a personality best suited to yours via pattern recognition; bots that can respond to esoteric queries no human could plan for by using predictive analytics; and bots that disregard irrelevant data to hone in on a client’s precise needs based on a neural network built from the entire history of interaction—this type of AI aims to mimic the way neurons work, and never stops learning.

An example of AI-powered bots in action is the AnzacLive chatbot Wizeline developed for News Corp Australia earlier this year. Our team of AI and machine learning engineers took the extensive diary entries of Australian soldier Archie Barwick, and programmed them into a bot that answered questions to curious Facebook users.

If you’re interested in learning more about how our bots team is building world-class products and delivering great customer experience for our partner, fill out this brief form.

Luisa Posted by Luisa on Monday, November 7, 2016.

Comments

Leave a Reply

Mar 23, 2017

Your email address will not be published.
Required fields are marked *

*

Brought to you by Wizeline LEARN MORE