categories

HOT TOPICS

NEWSLETTER

If you are considering becoming a 1M/1M premium member and would like to join our mailing list to receive ongoing information, please sign up here.

Subscribe to our Feed

Thought Leaders in Artificial Intelligence: Andreas Wieweg, CTO and Andy Peart, CMO of Artificial Solutions (Part 2)

Posted on Tuesday, Apr 12th 2016

Sramana Mitra: I’d like to understand what customers you are going after and what the use cases are in those customers.

Andy Peart: Since it’s a platform, we, our partners, or indeed our clients can use it to build a range of different point solutions. We would refer to something like Siri as something of an artefact.

Indeed, we have an equivalent of Siri that we showcase. It’s called Indigo. It’s available across multiple platforms. It’s used by billions of people now. It’s getting rave reviews on app stores. This is an example of what we can build, but that wasn’t really the example that I was wanting to talk about. Let me bring it to light and put some real names behind it.

Shell is an interesting example. It’s interesting because they want to improve their customer service to users of their oils and lubricants. Using their own people as well, we were able to rapidly build a number of digital assistants that have a vast amount of information because they need to understand over 10,000 different Shell products across 130 different countries in multiple languages. They need to be able to identify and distribute relevant data sheets and there are over a hundred thousand data sheets. They need to understand 30,000 competitor products and 2,000 obsolete Shell products. They need to do this for over a million different engines.

One simple example is, “What oil would I need for my General Motors?” It’s a good example but where it really comes to life is when you’re providing lubricants to $5 million earth-moving machines where, if something goes wrong, Shell is going to be looking at a loss of liable money. They have to make sure that it’s absolutely correct each time.

Compounding the difficulty of all of these, the product names are pretty complex in their own right. They tend to be made up of a word that sometimes is not a proper word, or message, and numbers. It’s a complex search. Then people ask things in multiple ways and they are going to need those queries qualified to really understand.

For example, it might be a distributor inquiring about a competitor product. What’s the competitor product to ABC? What’s the Shell product to competitor’s product ABC? It would be able to ask more information. What particular application are you looking for? Are there any specific characteristics? It will, in a humanlike way, ask relevant questions until it has enough information to be able to provide the right answer there and then.

Additionally, you can’t program to support those multiple millions of different connotations so you need an intelligent system that is able to understand the multitude of different ways that a customer can ask a question to be able to access the relevant information from a corpus of knowledge, pull back that information, and incorporate it in the answer, and then listen so that you can start to get trends. That’s one example of how it can be used.

This segment is part 2 in the series : Thought Leaders in Artificial Intelligence: Andreas Wieweg, CTO and Andy Peart, CMO of Artificial Solutions
1 2 3 4 5

Hacker News
() Comments

Featured Videos