CEO and Co-Founder of ViSenze
Oliver Tan is the CEO and Co-Founder of ViSenze, a visual search recognition company that is transforming commerce through Artificial Intelligence. Prior to starting ViSenze and running another startup, Oliver worked in the corporate world for over 20 years, but eventually decided that he wanted to make more of an impact in the world. At the age of 44, he left his corporate position and rejoined the tech world and he is now full immersed in AI.
ViSenze powers visual commerce at scale for retailers and publishers. The company delivers intelligent image recognition solutions that shorten the path to action as consumers search and discover on the visual web. Retailers use ViSenze to convert images into immediate product search opportunities, improving conversion rates. Media companies use ViSenze to turn any image or video into an engagement opportunity, driving incremental revenue. The company was recently named an Interbrand Breakthrough Brand 2017, and is consistently recognized as a global leader in the visual commerce field.
What triggered the idea of starting ViSenze both from a personal and business perspective?
For me, it was really about making a difference that could significantly transform lives. I felt that I wasn't making a big enough impact in the corporate world. We had some successes, but when I looked at myself heading towards retirement, that's not what I wanted. So I got in touch with my current co-founder who I’ve known for 12 years, and we began looking at this technology that was being developed at the National University of Singapore.
At first people challenged us and said, “What is a visual search for eCommerce? Doesn’t Google already do that?” Google does a great job with searches, but we wanted to develop a really smart tool that understands the way you shop in the real world and bring that online. People search for products online because they don't have hours to spend in the mall, so if we can accelerate that process to make it more efficient, more discoverable and more positive, we then have an opportunity to accelerate buying behavior.
What convinced you on a personal level to take the risk of starting ViSenze?
It’s quite uncommon in Singapore for well-paid executives to leave their comfort zone and jump into something totally new and unknown, no matter how adventurous it is. But for me it was not all about the compensation. It was about wondering whether or not I wanted to do a job in a suit for the rest of my life. I had always wanted to do something in tech, but never really found the right time or the right place. I also realised that if I didn't do this, somebody else would, and I would end up kicking myself for the rest of my life for not making the effort and taking that risk. So I made my decision and haven’t looked back since.
Can you tell us a bit more about what ViSenze does?
ViSenze combines machine learning and computer vision to create product search applications in retail. We enable product searches with what we call a shoppable lens that is designed specifically for a product search, without the use of keywords. Once this is embedded inside your phone, it becomes a shopping tool. When you fire up the camera lens, you can use it to find products from companies like Zara, Zalora, Uniqlo, and Amazon. This is what I call an omnipresent visual commerce experience. Think of us as a democratized version of Pinterest at scale on every platform. Our job is to make every image shoppable, whenever and wherever the consumer journey starts.
What trends or technology do you think will shape online consumer behavior and the new retail experience?
One of the biggest changes driving retail today is the change in consumer habits. All of us behave differently. Today, studies show that three in four consumers are inspired to make purchases by image and video content seen across retail and social media sites. Social media as well as the increasing amount of visual content available online influences consumer purchase behavior more every day. For example, millennials’ discovery and search patterns are primarily driven by social media content. They go on social media to get ideas, then they go shop online. These kinds of changes in habits drive retailers to stay on top of their game and implement new strategies to reach out and engage consumers.
What are your thoughts on the direction AI is taking?
In the next three to five years, I see AI taking form and shape in two ways. One, is what I call broad AI and the other is called vertical AI. Broad AI would be like Siri for instance. If I were to ask Siri, "Find me the meaning of obnoxious," Siri could give me the meaning. If I were to ask, "Siri, find me the best Ukrainian beer bar here in Singapore,” Siri wouldn’t know where to look for that because Siri doesn't know the hyper content that exists behind it. So we need vertical AI agents.
Do you have any concerns about AI?
One of my concerns about AI has less to do with the technology itself and more to do with the responsible development of AI and creating positive net impact to our daily lives. We should definitely not fear AI but neither should we be complacent about the dangers of not setting guard-rails as we rapidly push forward.
We used to think that we were the most intelligent species on this planet, but now we are teaching machines to be more intelligent than us, and there are many people warning us about the dangers. So either we’re the smartest species on this planet or the dumbest because we are about to create a super intelligence that can overtake us.
Another concern I have is whether or not we are doing this in a collaborative and coordinated way or if there is a rogue AI project somewhere in the world teaching machines to do specific things with sinister objectives. There are things that we are not aware of that could potentially break the boundaries of ethics, safety and privacy, and these are points we all know that we should not go beyond.
The Cambridge University Centre for the Study of Existential Risk puts out a report every year highlighting the 10 biggest threats to mankind, and this year was the first time ever that AI was included in that list alongside nuclear war, asteroid impact and food shortages. So our awareness has now risen to the level that we certainly need to be mindful of it.
What particular challenges are we currently facing with AI?
One of the biggest challenges currently faced with AI is the constant need for learning and training with data for machines to perform better, especially in unstructured environments.
There are definitely some mistakes and failures that AI has encountered. Take Tesla for example with their autonomous driving mode. There was a very tragic accident where a gentleman took his hands off the wheel (although Tesla specifically does not encourage that) and the image sensor could not discern between a white trailer truck and the sun shining directly on the vehicle. It just goes to show that we should not over-rely on AI.
Another example happened just today when I was asked to interpret images related to violence. If I look at an image of someone putting a knife to someone’s throat, is that violence? Let's say the people in the scene are just having fun and one person waves a knife at another person while smiling? Is that a violent scene? What if it was a violent scene and AI was supposed to detect it on a CCTV camera, but interpreted the scene as two kids playing and having fun? So you can see the difference between human based experiential learning and machine based learning and the consequences of a real-world application of AI where AI could potentially interpret it in the wrong way. So, these are things that the AI world are still trying to grapple with.
What are some key lessons you’ve learned running Visenze?
We have a very open culture that tries to understand and assimilate the best technology and the best market understanding out there, and we have gotten better over time. I have this approach with my team that we should fail fast, fail small, learn big and learn deeply. If we have to make mistakes, we learn from them and amplify the learning so we can take that to new markets quickly. Another lesson we have learned is to be open to collaboration and learn from smart people out there too.
What books are you currently reading?
Right now I’m reading Super Intelligence: Dangers, Pathways and Strategies by Nick Bostrom. In it he warns about the potential pitfalls of super intelligence if we let it go uncontrolled. As a technologist, I used to think that any advancement we make is good, but very few people warn us about the potential pitfalls.
What are your go-to resources to stay on top of trends?
For hard data I mainly go to TechCrunch and Crunchbase. I also read a lot on Forbes. I'm a member of the Forbes Technology council in the US, so I contribute as well. For TED Talks I like How Great Leaders Inspire Action by Simon Sinek. He's an excellent motivational speaker. I’ll always remember the line where he says, "People don't buy how you do it. People buy why you do it. And what you do reflects the belief in what you are." A lot of people can explain what a product can do, but very few people can explain why it exists.
What are you going to learn in the year ahead?
I want to learn some basic coding. It’s ironic that I work at an AI company and still need to learn to code. I also want to learn more about the art of better persuasion. I think we can get a lot further if we win an argument with persuasion rather than logic. Yes I'm in data, and you can win an argument with data alone, but persuasion is actually the real art of winning somebody over. I think as a person I can get better at that.