Sramana Mitra: Can you take us through some use cases on what AI is able to do to the customer issues that you’re dealing with?
James Markarian: We’re a horizontal technology. In the integration world, you start out connecting one thing to another. By the nature of the fact that we are a cloud technology, we’re able to detect the patterns that our previous customers have used in order to connect one system to another.
If I say I want to work in the domain of human capital management and I’m migrating from PeopleSoft to Workday. I start by saying, “I want to move this employee data from one system to another.” As soon as you say, “I have a PeopleSoft employee object,” you can predict based on everything that’s happened on the product before.
Since we see these patterns so many times, we do a very good job of predicting what you’re going to do next once you start with a particular object. Because we’re horizontal, that applies to human capital management, CRM, and data warehousing. They’re all different, especially data and application integration cases, but the patterns are largely the same.
Sramana Mitra: What is the AI angle to it? I understand SnapLogic’s business pretty well from an integration point of view.
James Markarian: Let me be a little more explicit. We’re building predictive models based on the development artifacts that we’re in constant track of, by nature of being a cloud technology. Every time, somebody does something in SnapLogic, we record all of the metadata. Even if the data execution might happen on-premise, we know the general patterns. We can train models on those patterns so that when we see you trying to do something, we can do a very good job predicting what the next step in that process is going to be.
In this case, we use machine learning to build those predictive models. We’re an integration technology provider; not an AI provider. Nobody has done this. In the integration space, the software spend alone is several billion dollars. Nobody in that several billion dollar category has applied AI to the process of automating integration. From an economic point of view, it’s certainly relevant.
Sramana Mitra: Tell me more about applying AI to automating integration.
James Markarian: I think there are three things your’e trying to achieve. Integration is generally the barrier to executing a particular business process. I want a business process implemented or changed like marketing lead generation to sales funnel. Normally, it gets shoved to IT. It goes to some sort of backlog and the fulfillment of that process might take days, weeks, months, or even years.