We caught up with one of our talented entrepreneurs, Gil Sever, Co-Founder and CEO of Applitools, to get his thoughts on what it’s like to start and run an Artificial Intelligence (AI) startup. He’s been working in the technology space for many years, starting in the military managing small and large technical projects. After the military, Gil went on to be VP of R&D for a startup, COO of a publicly traded company, and then founded three startups in the fields of Cellular, Security, and Storage. Two of the three companies got acquired (by IBM and Wave Systems).
Check out his tips for entrepreneurs and insights into what it’s like to run an AI startup.
How did you decide to start an Artificial Intelligence company?
Two of my former colleagues (Adam Carmi and Moshe Milman) and I decided that we wanted to create a company and continue to work together. Adam suggested that we solve a pain-point that no other solution has solved – harness the power of AI to mimic the human eye and brain. This led to us creating what we decided to call “Visual AI”. We felt that with Visual AI we can change the way software is tested around the world.
What is Applitools and what products and services do you provide?
Applitools invented visual AI. We developed AI algorithms that mimic the human eye and brain and allow a computer to check web and mobile screens that could otherwise only be analyzed manually. Our AI algorithms are able to analyze visual interfaces much faster and more accurately than humans. We are a SaaS company and we are deployed in the Cloud.
The first application of our Visual AI is in software testing, making sure that web and mobile apps look the same globally on all different browsers, devices, screen sizes, and resolutions. Recently we added a Cloud testing Infrastructure that we call Ultrafast Grid, that allows users to render application screens on multiple browsers and devices in parallel and accelerate cross-browser testing and cross-device testing by 30x. With this service, we now provide a full end-to-end functional and visual testing platform which is better and faster than any other platform in the market.
What’s your favorite part of working at a startup?
There are many parts I love about working in a startup!
- I love watching my ideas become a reality. Initially, ideas are translated into slides and Excel sheets, then design documents and code, and then products that are actually used and loved by customers that are willing to pay for them and can’t imagine life without them. Working at a startup allows you to see this process of an idea ultimately becoming a product that can be used by tens of thousands (and sometimes millions) of consumers and Global 1000 companies. It’s an amazing feeling.
- You meet new challenges every day. These challenges change and evolve as the startup evolves. The things you do in the first month are totally different from the things that you need to do in the second year, and in the fifth year.
- You are not a small part in a big machine, you are the machine! And you get to view all the gears working together. You see R&D, Marketing, Sales, Business Development, Finance, and more. You get a view of all aspects of a company and you need to manage it all simultaneously.
- You are likely to be exploring new territories that no one else has explored before you. You are the first person in the world to see new technologies, to solve use cases that haven’t been solved before. It’s like a roller coaster for adults!
What’s the hardest part about working in a startup?
The hardest part is derived from the good part – it’s totally unpredictable. You can’t set your schedule in advance. For example, you can’t say, “I’m going to do this next week and then take a vacation!” You always need to be on full alert, be ready for something new to come your way. You need to be ready to jump all-in in a heartbeat, for as much time as needed. This can mean being available 24/7 for very long periods, where the startup is your top priority and some of your personal priorities like family and health for example may have to be deprioritized.
What challenges do you believe are unique to AI startups vs others?
AI is a new territory. It started a few decades ago but now it is becoming more mature, scalable, and effective. However, it is still rapidly changing. In addition to other innovations and surprises you have in standard startups, an AI-based startup is basically a “startup square” in terms of the surprises and the things you need to do. Because it’s a new territory, you actually don’t know what to expect. Developing AI is different from developing new versions of legacy software or hardware. Testing AI is also different from testing legacy systems.
There are so many new things you need to know and understand like:
- How to make AI self-learn and improve
- How to collect and analyze data sets that you can feed to your AI algorithms or to your deep learning process
- Which AI technology suits best the problem you are trying to solve – Neural Networks? Machine Learning? Deep Learning? Or maybe something else?
- What kind of hardware or infrastructure you need
AI adds complexity to the solution you design and implement, but at the same time, it can provide you with exponential improvements to your solution. AI can improve your outcome by 10x, 50x, and even 100x if you do it correctly.
Are AI startups different from others, if so, how?
AI startups can differ from other startups because the Research, Development, Testing, and even the outcomes are totally different. When you develop AI algorithms there’s the traditional part of it, like writing code and choosing hardware, but at the same time, there’s a new dimension of auto-learning and auto-improvement that is inherent to AI algorithms. That also requires different sets of skills like data analysis, machine learning, and more.
What do you see as keys to success in the AI business and the startup world in general?
In the startup world in general, you need to be curious, looking for innovation, and looking to achieve new goals that haven’t been achieved before. You need to be very persistent, never give up, think outside the box, and create the impossible.
To be successful in an AI startup specifically, you need to find a way to answer questions that can’t be answered effectively with traditional tools and algorithms. You need to know how to match the right AI algorithms or implementation methods to the specific problem you are trying to solve.
When you’re hitting a brick wall your first instinct should be – I must get through it! You should never give up because you can never know how close you are to success. You should always be optimistic, keep fighting, look for other ways to succeed, and eventually, you will win.
What advice would you give to other entrepreneurs working in the world of AI?
Stay very curious and continue to learn all the time. The frameworks, tools, and methodologies that are required to excel and to find new solutions in the AI space are rapidly and constantly evolving. You need to stay at the forefront of it if you want to succeed!
Are you fundraising for your AI startup? Check out these tips on How to Get Funding for Your AI Business.
Learn more about What Investors are Looking for in AI/ML Businesses.