AI stands for Artificial Intelligence. AI is more than a marketing buzzword and it is important that customer experience managers can truly understand its purpose to gain the maximum. Here at Sentisum, AI could also stand for Axel’s Interest. Axel Almquist is our NLP and Machine Learning specialist who spends his days customising and optimising Sentisum for individual clients. There is no one better to decode what AI is in customer experience is.
So, Axel, what is AI?
Well, AI is a super broad concept. In its broadest sense, I'd say it just refers to a machine who performs an intelligent service which means that anything from a calculator to a robot arm or a Tesla is an AI. But most commonly I think people think of something like the singularity when they hear AI.
Wait, what is a singularity?
Simply put the singularity refers to when technology (in some form) comes to a point where it surpasses us in general intelligence and becomes hyper-intelligent; meaning it can do everything we can do and more, something in line with HAL 9000 in 2001: A Space Odyssey.
So what AI is SentiSum?
Our AI, as most AI, is aimed to perform a task that we otherwise would rely on a human to do, with the goal of minimising error or improving efficiency.
Our AI, just as humans, can read and extract useful information from the text but many times faster. Our AI uses something called Machine Learning which enables it to learn from data that we feed it.
Why is everyone in Customer Experience so obsessed with AI then?
So, as stated, AI is often applied to improve the efficiency or performance of something we humans already do. Many companies today rely on manual means to e.g. understand what their customers like or not, which is very slow and costly. If we can make machines that can read and understand customer conversations x times faster than humans, we can not only save companies money but make sure that all customers are heard.
Do you then need to be an expert in AI to use it?
Well, you use a calculator, don’t you?
But to be honest, if you make an AI product that's supposed to help humans in any way and is hard to use, it's just a bad AI product.
What is a usable, good AI platform?
An AI platform is good and usable only if it works on your data, is simple to use and solves your specific problem(s).
Hence customization is a key but this could be difficult depending on the task and availability of data.
You mentioned Machine Learning, Deep Learning and NLP, what exactly are they?
In simple terms, Machine Learning is the ability to automatically learn from data and make decisions without being explicitly programmed(rules).
Deep Learning is a type of Machine Learning approach that makes use of Neural Networks. It’s sometimes also argued that it is only deep learning if you use Neural Networks of certain depths.
NLP (Natural Language Processing) is a sub-field of AI that tries to bridge the gap between computers and natural language. This includes everything from making the computer understand what a "word" or a "phoneme" is to detect emotion, write jokes or answer questions.
So, what do you do then on a regular day as a data scientist in NLP working on AI?
Most days are a mix of researching improvements for current solution, coming up with new ones, being deep into the world of coding or training our machine learning models.
Is there anything frustrating about your job?
Data. Deep Learning bugs.
What are the best parts of your job?
NLP is a difficult and relatively new field which still pose a lot of fun and challenging problems. So, the best parts of my job are designing and improving solutions to existing or new NLP problems we face.
How did you get into AI?
After my BA in Theoretical Philosophy I found myself in a cross road and decided to try something else than philosophy before pursuing an academic career. Unbeknownst to me I found something exciting and enthralling in the world of NLP and Machine Learning and I've never looked back since.
So, you weren’t learning to code as a kid?
No, I started learning when I was around 21. I think what really got me hooked was the instant and simple confirmation coding gives you — if it runs its correct (most of the time... ) — which is something you very seldom get working with Philosophy.
It is never too late to start learning about AI and NLP. If you want to learn then you should start whenever.
Finally, my last question I’ve always wanted to ask someone in AI is, will machines take over the world?
Haven’t they already?
AI isn't just for computer experts and technical geniuses. Good AI is accessible to everyone ! If you want to find out how you can start using AI then schedule a demo with us to see if SentiSum can help you.