Sramana Mitra: When you go into a direct account to sell, are you going into the CIO’s office or are you going to the business departments where you see the use cases.
Luca Scagliarini: The value of what we do even if it’s highly technical is perceived at the business level. The upstream group in an oil and gas company understands the complexity of processing unstructured and structured information together. It’s the security officer in some companies. It’s the compliance officer in certain companies. We usually go direct to the business. We try to limit the number of use cases because we cannot cover everything, but the use case is very broad. The CIO or IT team is the operational part. We have all the answers for the CIO becuase we’ve been doing this for many years in many different implementations, but we go to the business to sell.
Sramana Mitra: You said you implement custom solutions based on your discussions and based on the use cases for that business. Does that mean that you provide the professional services related to that or is that done internally?
Luca Scagliarini: We have two main situations. First of all, the pure software integration is not that complicated. In certain scenarios, it can be done also through a cloud API. The customization usually happens at the knowledge level. If I want to categorize my content, I might have a taxonomy that I consider relevant for me, which could be different for another company. It’s more of a knowledge customization. If I’m in the intelligence business, I want to extract all the information related to critical infrastructure. If I’m in the electronics business, I want to extract all the information about the product.
There is a level of customization that depends on the use case. That can be done by us, our partners, or by the end customers. The end customer can build, believe it or not, lots of intellectual properties. Being strategic about how you handle your information can make a difference. Some customers want to bring that knowledge in-house and they don’t want to share with anybody. We can train them and they can do the customization internally.
Sramana Mitra: I’m going to ask you more industry-level questions. You’ve said several times that the industry is still at its infancy in terms of how deep the AI applications are in the various kinds of use cases. You can be as technical as you want. Where is the industry and where are you trying to take it from an AI algorithm point of view?
Luca Scagliarini: Our vision didn’t really change from the beginning. We want to help machines to read as well as a person can. The first attempt was just to see if the sequence of characters is present or not. Then we started to do some enhancement towards having this deeper understanding. There’s still room there.