Sramana Mitra: We’ve actually covered several companies that do this both from a pure recommendation point of view. There are two very successful companies in content marketing that are using this use case to actually help publishers monetize as well.
Luca Scagliarini: That’s one case. The difference of using semantic versus a deeper understanding of content is that the level of granularity that you can achieve is much higher than what you can do with statistic algorithms. I really suggest people who are familiar with this approach to understand where you can go. If you think about reading all the content and not just thinking about the frequency of keywords, you can get a very granular user experience. That’s where the next step for publishing is.
I always make the case that we are in the first phase where people are taking advantage of that, but it’s not really what people would like to do. I would like to have someone who has all the content available and is providing me what is relevant for me at that moment in time. Only by having a deep understanding of that can you do it. I think we are just in the first phase. We can do much more.
Sramana Mitra: Let’s do your oil and gas use case. What problem are you solving for them?
Luca Scagliarini: There are several use cases in oil and gas. Let’s consider for example the situation where you need to optimize your production in your oil well. Usually, you have sensors that are giving you huge amounts of data. The reality is that the work of the geologist, which is actually the person in charge, cannot be based only on numbers. There is a lot of information that are coming out of reports. It’s information that has been developed over the years.
There’s a huge amount of content that you need to sift through to integrate these numbers that are coming out. That’s one use case of being able to weed and have a very high precision and accuracy in extracting the additional information that can help you in making your production better.
Sramana Mitra: Give me the context in which you’re doing this for oil and gas companies. It sounds very generic but I’m trying to understand where specifically are you applying this technology.
Luca Scagliarini: It’s in the exploration phase—exploration and production, which is the upstream part of the process. The main activity is to make sure that you optimize the production of your wells. Optimize means having the best productivity for your well. All this requires a combination of data and unstructured information. The unstructured information piece is where we help so that the geologist can make sure that they have all the most important information available in the moment that they work on it. This is a very vertical use case. It’s a typical situation of integrating data with information extracted from any kind of unstructured source.