Mr. Vrabel, your company claims to take the pulse of the planet by turning big data intoactionable intelligence – how did you come up with the idea and where did you start?
Frantisek Vrabel (FV): Believe me or not, but it started back in 2005. At that time, my previous startup I2S served on major programs in the framework of US Foreign Military Assistance. After achieving our mission, which was to assist in the success of the NATO enlargements in 1999 and 2004, I was thinking of what to do next.
We as a team had accumulated pretty extensive experience on so called C4I (Command, Control, Communications, Computers, and Intelligence), which is a lot about complex information systems, Big Data, the Internet and knowledge discovery.
In this context, I got the idea on creating the current system, which would provide actionable intelligence to multinational corporations operating on a global scale. We had completed the first version of the system at the beginning of 2008 already, but were hit by the global financial crisis later that year. It took several years for markets to recover and start feeling the need for our new kind of services.
What kind of help did SAP Startup Focus provide?
FV: SAP Startup Focus helped us with go-to-market, which was of critical importance for us. It has assisted us in getting visibility within the global SAP organization. We’ve been invited to present and explain the nature of our innovative data services on numerous high-profile SAP events, which were also attended by SAP customers. This way, SAP Startup Focus helped Semantic Visions with educating the market.
You say you broke to the code to distinguish critical signals from the noise in the background. How does that work?
FV: We automatically collect, analyze and synthetize 90% of world’s online news content, which is over 1 million unique articles per day. This way, Semantic Visions leverages the Long tail of Internet and is able to effectively run its early warning systems.
Semantic Visions is not a search engine, but a detection engine. There is a principal difference between the two capabilities, which many people don’t realize. You can hardly use Google for searching for the unknown, but you can use Semantic Visions productively for detecting events you’re not aware of.
It is vital to include as many sources as possible to increase the chance that the matter of interest is detected when it happens. But the more information you work with, the higher entropy you face.
Distinguishing critical signals from irrelevant noise is achieved in multiple stages. The predominant one is the semantic analysis in which precision, granularity and powerful multi-dimensionality play essential roles. The final stage is achieved in a process of so called Big Data Semantics, which is “nothing else” but the semantic analysis on multi-document and multi-language levels. Easy to say, but difficult to implement
Where does Hana come into play in that?
FV: Big Data Semantics is exactly the stage where Hana comes into play. Let me be specific on this. The preceding stage of semantic analysis is conducted by our own unsurpassed technology and without using Hana, because existing Hana text analytics is not applicable for this kind of task.
For Big Data Semantics, we use the unprecedented speed of Hana, where our sophisticated algorithms work with high-added value metadata (or Smart Data if you like) – the results of the semantic analysis stage. We squeeze the information “stuff” on and on until we get the sort of juice we need – the actionable intelligence relating to millions of companies. We do that on continuous basis and in real-time.
If you could pick one thing that makes Semantic Visions unique, what would it be?
FV: Well, the uniqueness resides in our entire solution. There’s no one on the market that is capable of providing effective risk detection in a wide spectrum of threats and relating to millions of companies and geolocations in real-time.
When going over your products, the terms “risks” and “early warning” come up a lot. How accurate are your predictions and what kind of data do you use?
FV: Technically, we don’t predict, but philosophically, we do. What is happening in the world, in most of the cases, we notice through media and other communication channels.
When Semantic Visions alerts you on a certain event, e.g. “Deteriorating financial situation” of one of your suppliers or “Emerging danger” in a geographical area of your interest, before you become aware of this through your current means, then we actually predict the future for you.
One has to take into account one more aspect. We as humans have rather limited capability to perceive the information, mainly if it’s too much of it. You can hardly keep a mental representation of tens of thousands of your suppliers, but computers can. Semantic Visions’ early warning system was designed to meet the needs of large corporations, to provide them with information environment awareness. Not to keep them blind, as they often are.
What are the customer-side requirements regarding IT-infrastructure and data availability?
FV: There are no special requirements other than being a customer of SAP Ariba. Semantic Visions’ solutions are fully integrated into this world’s largest business commerce network. As for Semantic Visions service to SAP Ariba, it’s DaaS.
The same model applies for the area of credit risk, where we work hand in hand with Deloitte.
From initial contact to having the customer run your product by themselves, how long does deployment usually take? How do you assist in getting your customer running?
FV: In this respect, we rely on SAP Ariba as our solution is integrated to it. Having said that, we as Semantic Visions start to deliver result data within 24-hour time frame from receiving the order from SAP Ariba.
With IoT and Big Data being the top trends in IT for the foreseeable future – what’s next for Semantic Visions?
FV: It’s definitely AI. We’re investing to this field big time.