In a previous blog post, you shared the inspiration behind founding SparkCognition. What has the company been working on since leading up to the recent Series B round of funding?
Entrepreneurship is often a balancing act between the practical aspects of running and scaling a business, and acknowledging the original aspirations that caused you to embark on your enterprise in the first place. In my case, it has been no different.
I see tremendous potential for artificial intelligence applications, such as our own unique Automated Model Building technology. It seeks to automate some of the work a human data scientist would manage, thus making AI applicable at broader scale with autonomously-built, self-maintaining models that free human talent up to solve other problems. However, I also know that you can't scatter shot your way to startup success, so focus - especially early on - is absolutely key. For the first two years after our launch, we were maniacally focused on developing a great footprint in the industrial and energy space, onboarding banner clients and solving big problems. The success our team has demonstrated in executing on this focus makes me tremendously proud - it has also been recognized by media and analysts alike. We recently ranked #20 on CNBC's Disruptor 50 list, which featured companies such as Uber, Airbnb and Palantir.
Over the last six months, we've begun the process of expanding into additional markets, beginning with finance and government/national security. Given our dedicated team and their astounding early success in signing up massive clients in both these areas, we’ll continue to demonstrate the efficacy of Automated Model Building and our other AI innovations in these new verticals.
SparkCognition contributed to Verizon’s 2017 Data Breach Investigations Report, which takes an in-depth look at the cybersecurity landscape. What are common threat patterns that your team identified in the report and how is SparkCognition leveraging artificial intelligence to enhance cybersecurity?
There are three major trends we are seeing. First, the mass proliferation of threats designed to defraud consumers and prey on individuals at large scale. Ransomware targeting individuals and small businesses is an example of this trend. Second, we see the weaponization and militarization of cybersecurity. Take the North Korean attack against Sony, as an example, or the attacks against the Ukrainian grid, and potentially the foreign interference in the recent U.S. election. Third, we see cybersecurity going physical in a big way. There is now credible evidence indicating that infrastructure providers - utilities, oil and gas companies, industrial systems - are being attacked in ways that may transform a cyber intrusion into a physical threat. The confluence of these three trends unfortunately make for a very dangerous environment.
SparkCognition is leveraging AI to enable cybersecurity in a number of ways. First, we are using a variety of models - including deep neural networks - to analyze the contents of files and determine if they are engaged in malevolent behavior. We’re essentially doing the job of a human security researcher when he or she statically analyzes a file. Our DeepArmor anti-malware product embodies this capability and can protect servers, virtual machines and a variety of endpoints. Second, we provide a set of cognitive pipelines which can analyze semi-structured information such as the logs generated by firewall or proxy systems. These pipelines "reason" over the information and find clues that firewalls and other blocking mechanisms have missed. By integrating with a SIEM (security information and event management) system, these pipelines can transform a log database into an active, cognitive repository that is looking for new threats and patterns and proactively flag them to your attention. Third, we have considerable Natural Language Processing (NLP) capabilities that can automate security research, evaluate threat hypotheses, reduce false positives, and augment the incident-response professional, among other security practitioners.
The company recently announced that its AI solutions are more readily available to government agencies via a strategic partnership with Summit Government Group. What role will cognitive technology play in addressing national security?
The role of cognitive technology in addressing national security cannot be overstated - it will be massively important. AI can bridge the skill set gap through cognitive systems and increase the speed of responsiveness through machine and software autonomy. It can also tackle the type of complex challenges that human brains are not optimally configured to address.
When it comes to national security, the impact will be both cyber and physical. We need to ensure we have free and fair elections, which is a cyber challenge. We need to ensure we maintain a military edge in a rapidly evolving world with complexity emanating due to competing and emerging power centers. That is a physical - or kinetic - challenge. AI will not only be crucial, but pivotal in addressing challenges in both of these dimensions.
SparkCognition has been well received in government circles and we are incredibly grateful to both the civilian and military leadership for this. We are now listed on the GSA (Global Supply Annual) catalog and are actively making our solutions available to government agencies and the military.
In your experience, what’s the best way to work with investors?
In my experience, I've found that if you truly believe in what you are trying to do and have faith in your team and your market(s), you're already well on your way in developing a message that will resonate well with good, experienced investors. Don't try to transform and alter your message to please or satisfy a specific investor. You'll find this will actually end up being counterproductive. Of course, if you speak with a large number of people and no one is interested, then you've got some soul searching to do and some adjustments to make. But in general, find investors who believe in what you're trying to do, and be direct and honest with how you intend to grow your business. All sophisticated investors understand that when you go to battle, the enemy gets a vote... plans change and evolve. The important thing is how proactive you are in gauging the situation and adapting and learning from success and failure.
Remember that investors are investing in you, your team and your business because they feel you have what it takes to succeed. They may not be experts in your field, and most certainly won't know anywhere near what you know about your own business. It’s important to know how and where to ask them for help. In general, I've found the strategic relationships and connections with the (right) investors' networks are of great value.
What’s next for SparkCognition?
Miles to go. We are at the very beginning of a cognitive revolution that will transform every sector of almost every industry, as well as impact every economy and government. Rewiring the world with AI at its core - which is what I passionately believe is happening - will be a multi-decade process and SparkCognition will be a key player in making it happen. If you think about the three areas we've focused on - industry/energy, finance and government - you'll realize that this selection wasn't tactical or driven by near-term thinking. We picked the three areas where we believe AI can have the greatest impact and where we can deliver truly meaningful solutions. This process will continue for us and will accelerate as we apply greater resources to drive even better results for our clients and, one customer at a time, build the world we want to live in.