4.5.23-AI-Sparkdit

FADI MICAELIAN: My name is Fadi Micaelian. And I'm going to talk to you about SPARKDIT. We have a cutting-edge AI that's going to revolutionize the shopping experience. We do it by personalization and recommendation without ever intruding on the consumer data. So the problem for our customers is that they have a dismal conversion rate. You go to their website, at best 3%, usually it's between 1% and 2%. In order to do-- in order to address this problem, they need to implement AI.

But the key competitors are Amazon and Google. And with Amazon and Google, they can't have better algorithm. They can't even get close to the amount of data that those other guys have. So what's the solution? The solution is our recommendation engine, the best recommendation engine there is. 2x better than anything else that's in the marketplace that leads to a 20% increase in sales. And that with zero private data. I say zero private data.

So you're going to ask me, how do you do that? I mean, without private data, how can you enhance the performance? So we've done a new AI technology. And that new AI technology is based on three pillars. The first one is we mimic the way the brain works. And that's a departure from traditional your grandfather AI technology. And we focused on, how do people think? People think by tradeoffs. So we built a tradeoff engine. It mimics the way the work-- the brain works.

The second thing is when we don't have enough data, we ask the users. The internet is by nature interactive. So why guess what they want? Ask them. Engage with them. And in sales, nothing is better than engaging your customer. The third thing is build trust. Nothing builds more trust than not stealing your customer data, not spying them-- on them, not stalking them. I don't like someone to stalk me in the mall. I don't like someone to stalk me on the internet.

How many times have you been to a website, done a search on cowboy boots, and then gone to CNN, and start seeing all those ads on cowboy boots? You're like, how did you know about that? Sorry. My mistake.

So as a result, we're getting 88% conversion prediction rate. That means we predict 88% the right product versus 49% to 45% in collaborative filtering, AI-based collaborative filtering, and 37% in content filtering. The secret behind-- the secret behind our technology is only one thing. Instead of just focusing on data element, what we're focusing on now in volume of data, we introduce the concept of knowledge, domain knowledge.

And by domain knowledge, you immediately-- the first question is you're not scalable if every time you need to bring an expert. That's where you use AI. We use AI to essentially get the domain knowledge and embed it into your information. And that's where we change the whole-- it's game changing.

I'm going to give you an example, if you go to the website of Nike, they'll ask you, is it Flyknit, Lunar [INAUDIBLE] or Flyease? And we went to a soccer field and asked about a hundred players, which one of these? Not a single one knew what those technology were. And that's the key question about in differentiating the product.

Instead, we went-- we used our engine and we told our engine, what do people use when they play soccer? And they came back, says, speed, dribble, shot, and precision. And so what we do, we ask the users, tell us about your playing style. And it's not one or the other. It's a balance. It's a tradeoff between those different skills. And that gets you a significantly better recommendation.

Example of people who use our technology include HR in the United Nations. FAO intervenes when there is a tsunami, or shortage of food, or water contamination. They need to bring the best resources. And that's when they use our technology. They trade off the expertise in that domain, the distance to the location, whether they are an employee or not, et cetera. Another one is procurement multi-criteria reverse auctions.

Today, reverse auctions are price based only. So you get the cheapest thing, not the best thing. Today, when you have multi-criteria reverse auction, you can bet-- get the best solution, not just the cheapest one.

The third one is emergency response rate. Anthrax spread, or you don't know what spread, you need to respond immediately. And the defense intelligence is using our technology to decide what to do. Do you quarantine people in a stadium? Do you tell them to stay home? What's the best route for the doctor? Who do they-- who should they see first?

The last one is the application of Boeing, which is a swapping application where they take, an airplane comes in, has a problem, whether someone dipped a coffee, or spilled coffee, or the engine is not working. How do you swap planes? And so they have a decision and they take multi criterion into effect.

What we're looking for is three things. One is retailer who are interested in doing-- being pilots or A/B-- doing A/B test. Two is people who have financial service expertise who want to do financial services. And the last one is in others places. Thank you for your attention.