Published on Image & Data Manager

Ike Kavas is the founder and CEO of Ephesoft, with 20 years of document capture, document management, workflow and systems engineer experience. His previous role was Chief Technology Officer at Ephesoft and he holds a patent on supervised machine learning for document classification and extraction. On a visit to Australia in May for a series of customer events, IDM asked Ike to expand on expand the company’s mission of turning the world’s unstructured content into actionable data.

IDM: You started the business in 2010 looking to revolutionize the enterprise capture. In your mind, what were some of the biggest pitfalls of this technology?

IK: I think there were three main drivers in my head. One is, I felt like the innovation was not there. The vendors at the time were just updating their UI and not adding any significant value to the space so I wanted to bring something new. The second one was I saw the cloud is coming. Number three was as a result of being so closely involved with many projects for content classification and extraction on unstructured documents. They would take so long, many many months and millions of dollars. I wanted to bring that technology to the masses and you can only accomplish that by making the product accessible and easy to implement. This is one of the reasons we brought in subscription pricing eight years ago. I love subscription because it makes us work harder for the customers because next year they can just switch to another vendor if we don’t keep up. So, it’s good for customer good for Ephesoft and keep us on our toes.

IDM: How do you believe Ephesoft differs from other capture solutions?

IK: I think number one is the licensing model. We are very flexible and we don’t charge for click count and number two is the innovation. My nightmare at night is that somebody is going to come and ‘out-innovate’ us like we did. That’s why we work really hard and our self-mandate that we put in this is that we need to innovate something new every two years, so we have been trying to do that. I think that’s our main differences so that constant improvement you see from Ephesoft.

IDM: Where does smart document capture end and business process begin, are they the same thing?

IK: That’s a good question. You have to think of us as data transformation service. We take unstructured content, whether email, document, pdf file whatever it is, and we turn that into information and give it to someone else. You can run your workflows with that and that’s where the difference is. We are about data transformation, so we turn down unstructured content into structured data and then the rest, the other business processes do it.

IDM: And how does that fit with something that’s getting a lot of airplay these days, robotic process automation (RPA)?

IK: I think the RPA is very popular because it allows you to access the data and combine it with other sources just to transform it. In RPA you can say ‘go to this website and capture this data’ or you could ‘go to this database and capture this data’. But you can’t say that ‘go to this pdf file and capture the data’. We are the layer between that pdf file and what RPA needs. We can go to repositories, pdf file or any unstructured content and turn that into structured data that RPA can then utilise and transform so we are an essential piece of the puzzle.

IDM: Thanks, that nice clear explanation. It’s been a long while we have been illustrating the fact that organisations are struggling with unstructured content by quoting the Gartner figure that it represents upward of 80% of all available content. Is that figure still valid, improving or getting worse – What do you think?

IK: Unstructured content is I think growing four times faster than the structured content. One of the reasons is that we have democratised content generation. For example, we now allow people to take a picture of documents and add a digital signature before submitting them. But every time you digitally sign a document, that turns into a pdf file. Guess what, that’s another document stored in your repository as unstructured content that we as Ephesoft need to turn into data for actionable information. It’s definitely getting worse and that’s why Ephesoft is going towards the cloud, you need hybrid cloud elasticity because the sheer scale of data is not something that you can handle by putting some more servers on-premise any more.

IDM: There is also a lot of hype today regarding AI, analytics and machine learning. Where do you think the main applications of this are coming?

IK: I look at the AI and machine learning from two angles. One is how I can drive down the professional services engagement so that the customers from Ephesoft can actually implement these technologies faster. Through machine learning and AI we can have the computers analyse the data and extract what you need rather than us or the business analyst writing rules or repeating the system where the data is handled template by template. Those are really old school things. The second portion is how we can empower the business analysts. When you look at the data, say you have 2 million records that you reconciled from all the documents that you had in that process and then you combine that with other databases, the business analyst has to make intelligent decisions and you need to make sure that you empower them to make the intelligent decisions faster. So, by analysing this data the business analysts can find the data they need and then make the decision that they are supposed to make.

AI is all-encompassing and machine learning has different flavours. We really believe in supervised machine learning because we believe in that AI is not yet smart enough to figure out what’s important for businesses analysts. You must tell the machine ‘these are the important things that I care about for my business for my process’ and computers need to learn and amplify that so that’s why supervised machine learning is the basis of all of our technologies.

IDM: How do you see the future of document capture and Ephesoft and from here on?

IK: Enabling cloud is critical. In a few months we plan on empowering our TransAct product with the flexibility of the cloud. Many of our customers struggle with their peak volumes as when you buy an on-prem system you have to buy X number of servers to support your peak volume so you can process on time and meet your sales. But with the cloud that’s not the case. You can actually have only one server on-premise and when the peak volume hits you can just outsource or do everything on the cloud. That’s the short term roadmap but in the long term we have to be the best of the best on transforming unstructured data and giving that structured version of the data to our partners and customers and. If we just focus on that we will not only create great value but also become a baseline system for all of our partners.