Your Reflection, but High-Tech – Building an AI-Powered Smart Mirror

For this episode of Art of the Kickstart, we caught up with Edward Zhong, the creator of Ayi, the AI-powered smart mirror. Tune in to learn more about how Zhong created a smart mirror that is not only compatible with thousands of apps, like Uber, but also is able to custom-fit every user by combining artificial intelligence (AI) with an everyday-use product.

Topics Discussed and Key Crowdfunding Takeaways

  • How to find useful, day-to-day integrations with AI
  • Hardware and software challenges Ayi faced with their product as a growing startup
  • Their best advice and tips for someone developing a product with AI
  • Things to take into consideration when directing and shooting a crowdfunding campaign video
  • How to manage crowdfunding feedback when over 60% of your backers are first-time crowdfunders



Art of the Kickstart is honored to be sponsored by The Gadget Flow, a product discovery platform that helps you discover, save, and buy awesome products. The Gadget Flow is the ultimate buyer’s guide for cool luxury gadgets and creative gifts. Click here to learn more and list your product – use coupon code ATOKK16 for 20% off!


View this episode's transcript

Roy Morejon: Welcome to Art of the Kickstart, your source for crowdfunding campaign success. I’m your host, Roy Morejon, President of Inventus Partners, the top, full-service turnkey product development and crowdfunding marketing agency in the world. We have helped startups raise over $100 million for our clients since 2010. Each week, I’ll interview a crowdfunding success story, an inspirational entrepreneur, or a business expert in order to help you take your startup to the next level with crowdfunding.
Art of the Kickstart is honored to be sponsored by BackerKit and The Gadget Flow. BackerKit makes software that crowdfunding project creators use to survey backers, organize data, and manage orders for fulfillment by automating your operations and helping you print and ship faster. The Gadget Flow is a product discovery platform that helps you discover, save, and buy awesome products. It is the ultimate buyer’s guide for cool luxury gadgets and creative gifts. Now, let’s get on with the show.
Welcome to another edition of Art of the Kickstart. Today I am talking with Edward Zhong, the creator of Ayi, the AI-powered smart mirror for your home. Edward, thanks so much for joining us today on Art of the Kickstart.
Edward Zhong: Thanks for having us, Roy.
Roy Morejon: So this technology is really cool. I mean, basically it’s meet your new personal assistant inside your home, this amazing, visually stunning AI-powered smart mirror. Many times, if you’ve stayed at these luxurious hotels and things like that, you’ve had the opportunity to play with it or mess with it, and now, you can bring this to the comfort of your own home.
This device launched a few days ago. It’s already over $80,000 in funding and hundreds of backers have already supported it. So Edward, I’m really excited to hear about where this all starts and what inspired you to create this smart mirror.
Edward Zhong: Yeah, sure. We really love our product, and we really hope that our customers will love it as well. This whole company started behind the vision of changing how the smart home industry fundamentally is really. Most of the members on our team are either former product designers or software engineers for smart home based companies like Nest or Google. Through our experience in this industry, it’s just seeing that, although there are really cool products being made every year, it seemed like the smart home industry was rather dull and a little bit boring.
We really wanted to change the status quo of the smart home industry in general in bringing better design, more functionality, and kind of a more futuristic feel really and take that risk that other big companies were hesitating to take for some reason. That’s where the vision started. Then, we started with the smart mirror because the mirror is an object that is very recognizable and already in everyone’s homes. Making that object, as you will, a very simple object that everyone already is familiar with into our first kind of flagship product seemed like a gut instinct move. Really for us, it was very natural to build out a lot of the features, and this is something that we wanted for ourselves for a long time. We’re very passionate about the product and the vision. Yeah, we hope everyone else, when we deliver the product later this year, will feel the same.
Roy Morejon: Oh, I’m sure they will. Let’s talk about the process. When you guys were creating Ayi, what was that process like? How did you go about deciding what features to include, potential stretch goals, or different integrations into the device itself?
Edward Zhong: Sure. That’s a great question because we feel that, in general, when people market a product as something that is AI-powered or really anything that has artificial intelligence or machine learning embedded into the marketing, it’s really a fad because they don’t understand, first… Many of these companies don’t understand what true artificial intelligence or machine learning is. They’re saying it as a buzzword so that if people who like technology or consumer tech will buy into it. But, the AI behind the actual product, it really doesn’t do anything to help the user. We wanted to steer that in a completely different direction and move in the direction that created apps that were actually useful for the user.
So when it comes to the usefulness, we kind of examined what someone would actually use our product for. The first thing that comes to mind is augmented reality. We started with a huge focus on AR because we know that this is the purest form of augmented reality you’re ever going to get. Of course, when you have an iPad or an iPhone and you open the camera app, you can have and implement basic augmented reality through the camera lens. But in a smart mirror, you can actually see, not only on a bigger display, but in a true natural environment, the clearest most crystal form of augmented reality.
This can be used for real estate agents who want to show potential clients potential furnishings and interior decorations in their new homes, and they don’t even have to stage the house anymore. They can open up their clients’ imagination, and hopefully sell more homes and make their clients happier this way. It can be used by actors and actresses who need to switch their makeup every morning so that they don’t need to go to a makeup store, Sephora, every single day and pick up products. They can try it right on their face using augmented reality in the Ayi smart mirror.
There are a lot of these unique, pure applications that we thought of that are not just marketing fads for cool tech. It’s real things that people need that can be embedded in just what happens to be a really cool product. We started with augmented reality, and kind of proceeded based off this mindset, based off this thought process of actually useful cases of AI instead of just full use of the AI and that’s this case with all our applications, all our integrations.
Roy Morejon: So what have been some of the biggest challenges that you guys have encountered while designing this product?
Edward Zhong: Sure. There have been a lot of challenges I would say. The biggest one, probably… Along the timeline, we started with hardware first. Hardware, the biggest challenge was really thinning it down. Now, we could’ve shipped with a five-centimeter CIC pacing, which is what the default setting was when we first started to develop it. That was really the size that our manufacturers were comfortable fitting all the electronic components into, along with the camera and all the sensors, but we weren’t happy with that because it doesn’t matter how many excuses there are, people don’t like a standing mirror and that will never change.
What we decided to try was to put all the core electronics into a separate compartment that is attached to the charging cord, and by doing this we were able to bypass the thickness, distinct thickness, and able to narrow that advice down to two centimeters, which we’re still trying to get down. Two centimeters allows most people to be able to mount it flush within their walls so you wouldn’t even notice there’s a protruding mirror there. That took a lot of time in terms of hardware.
The second and probably biggest challenge we had on the software side and that we’re still continuing to face to be honest and still like working on is gathering the data for the machine learning because, as a startup, we are not Google and we don’t have access to their massive machine learning database that they have when they develop a new AI application or a new deep learning application.
They have billions of people with user data to pull from, and they can learn from that and they can make judgments or presumptions based off all that data. We don’t have billions of users, and we cannot use a big company’s user data. So what we have to do is basically pull from as many data points as we can through… Kind of like just scrapping together data [inaudible 00:08:46] if you will and doing data cleaning, so making sure we get rid of data that is outlying or won’t help the algorithms and from there, kind of do the groundwork that all these big tech companies are doing to build the apps. So that’s a really difficult aspect that has nothing to do with how good our developer is and nothing to do with how hard we’re working. It’s just something that is the way it is, unfortunately, and we have to just work 10 times as hard as they do to achieve the same result.
Roy Morejon: Yeah, I’m sure. So what tips would you have, then, Edward for someone else looking to develop a techie product like this?
Edward Zhong: I would say if someone is trying to develop something with an AI app, my biggest tip would be to use as many open APIs and already built things as possible because most creators tend to underestimate the amount of work that goes into creating something custom whether it be an operating system or even just one app. It takes a long time depending on the complexity of the app whether it be for Android or iOS or any other platform.
I think that if people tried to do their research and adjusted their products to things that can be built without too much customization, it would be a lot easier. Of course, if there is absolutely nothing out there that is not the feature that a creator wants to build, then there really is no choice but to custom build it. Then, I think the evaluation has to be made, like does this feature really make a groundbreaking difference that we have to really go in and custom build it. Then once you know… Once you make that decision, it’s kind of like [inaudible 00:10:35] from there. But yeah, using as many things that exist as a foundation would be definitely be my advice.
Roy Morejon: Absolutely, solid advice there, Edward. So let’s jump into the crowdfunding campaign and specifically on your campaign video. I really liked the video in terms of showcasing all the use case and then highlighting a founding member of the team in terms of the product itself. What did that process look like, and how did you guys go about deciding what to include in your video, and then what got left out of the video?
Edward Zhong: Sure. This is a long story. We had at least actually four or five different videos created throughout the preparation of launching a product. The reason it was so difficult for us to do any sort of asset creation like photography or video was because a mirror is just physically probably the most difficult thing to photograph because when you… It has to do with the lighting and reflective surface of it, you can’t obviously have the camera in the shot, a lot of small technical things that add up and make task really difficult. We did a lot of different shoots with different people in the studio before we got got everything actually together.
One of the versions… How we decided the final video was really seeing the target audience respond on the live campaign, and we realized a couple of things, right? First, before the campaign, you know we are mostly on the younger side in terms of just the average age of our team so the taste preference of our own personal preference is going to differ vastly from what kicks better with the actual target audience. I would probably venture to say that the people who are purchasing our product are, on average, double our age, so we had to actually switch it up the third day of the campaign.
We switched the music. We basically switched the speed of the video a little bit, make it slower to match that target audience. I guess the biggest difference was that they preferred a more refined, slower, and less jumpy [inaudible 00:12:53] as with everyone we had validated who with what was like young like ourselves and wanted a more hype, you know, Apple iPhone X kind of feeling to it. We really should’ve done a little more targeted validation before, but it’s good that we made the switch because we’re seeing an okay conversion from it currently.
I would advise probably when people make the video going forward to not rely too much on personal preference even though that is important in the long run, but also to think about people on Kickstarter in general like.
Roy Morejon: Yeah, no really good insights into there. In keeping in that thread, let’s talk a little bit about your experience with your backers so far. Now that you’ve had a few hundred backers in there, it looks like you’re getting some really good feedback. How have you been going about managing the feedback and promoting the campaign? But also looking at your backers, over 60% of the backers are first-time backers on Kickstarter. Has there been any insights into that in terms of the community that you’re starting to build?
Edward Zhong: Yeah. That is really interesting, and actually, we don’t know why, but 60% of the backers are first time. I think a lot of it is probably due to our network and friends during just the first day, us telling them to come in. Most of them have never heard of Kickstarter before and created an account. But, you know, research on how to build a community is important for the sake of the product. What we have to focus on mostly is balancing that feedback we get with actual product development and realisticness because what most of our backers, unfortunately, are not going to understand is the complexity of the product. When they ask for a certain thing, they may think it’s easy, but it is in reality not. We also have to be able to manage expectations.
For example, just as a very tangible example, when we asked for stretch goal suggestions, a lot of people reached out and just gave a lot of suggestions. For example, like can you make something… Can you make a separate operating system exactly like the Tony Stark Iron Man one that looks like the one in the movie and that performs like Jarvis? Well, we can’t because an entire operating system like that takes like months. The other reason is that, of course, the movie operating system and UI doesn’t actually work. If you actually were to build that, it would be extremely unfriendly to users, and you wouldn’t even know how to use it. It would just look cool, and that’s about it.
There’s some things like that where you really have to be stern as a creator, and backers will respect it because they respect you, they know, you know… You’ve basically shown that you know what you’re talking about and you’re not just agreeing to random things to make them happy. That’s important for us. But, yeah, it’s great to see the outpour of support and we still value their feedback a lot. We just have to keep a good balance.
Roy Morejon: Absolutely. What’s been the biggest surprise for you so far early on in this campaign?
Edward Zhong: The biggest surprise was probably the fact that we had to… Well, it was probably the trend of the campaign in general. When you look at Kick Tracks and like relatively successful projects, it goes up on day one and then the trend is like down, down, down, down, down, and the towards the end, it might go back up. For us, the trend is going to look really weird, and it’s not a good or bad thing. It’s just going to look weird. It’s like up and then down and then it goes climbing back up, which I think is cool because I’ve never seen anything like it. I looked at a lot of Kick Tracks doing analyzation. That’s probably been the most surprising part. I don’t know what to make of it really. It was just surprising to me.
Roy Morejon: Awesome. Well, Edward, yours is going to get us into our launch round where I’m gonna rapid fire a handful of questions at you. You good to go?
Edward Zhong: Yeah, sure.
Roy Morejon: What inspired you to be an entrepreneur?
Edward Zhong: I don’t think anything really inspired me. I think it’s just in my nature.
Roy Morejon: If you could meet with any entrepreneur throughout history who would it be?
Edward Zhong: Probably having an espresso with Steve Jobs, either Steve Jobs or Wozniak would be really cool.
Roy Morejon: Nice. What would be your first question for Steve or Woz?
Edward Zhong: If you could take more free time… Well, I know for Jobs because he’s passed away, I would ask him if you could take more free time during your living life, would you have done it?
Roy Morejon: Yeah. What about for Woz?
Edward Zhong: Well, he’s still living so I don’t know.
Roy Morejon: Fair enough. We won’t bring him, then. What book would you recommend to our listeners?
Edward Zhong: I’m going to be very honest. I do not read a lot of books so I would not [inaudible 00:18:10].
Roy Morejon: How about a blog?
Edward Zhong: A blog, that I can help with. If you’re creating a Kickstarter campaign, I would go on the Kickstarter forum. It’s not necessarily a blog, it’s a forum, but it’s a pretty good resource.
Roy Morejon: Absolutely. Where do you see yourself in five years, Edward?
Edward Zhong: I just see myself continuing to make products, making great products, and yeah, that’s what I love doing so that’s what I see myself doing.
Roy Morejon: All right, last question, Edward. What does the future of crowdfunding look like?
Edward Zhong: Oh, wow. The future of crowdfunding I think… Well, two directions. I think one is equities crowdfunding is going to be huge because I think people love even bigger ideas. I’ve seen, you know, private jet companies start doing like, what do you call it, a cryptocurrency-based equity crowdfunding. I also think crowdfunding is going to be a great way for younger and younger entrepreneurs, in fact, to get into actual hardware. I personally love consumer tech and hardware, and I think it’s just something tangible you can like hold in your hand instead of software where you code it, it goes somewhere online, and it’s just stored in some sublime or text editor. I really hope that crowdfunding is an actual inspiration for more people to get into hardware. I think it’ll inspire more young people to be able to create tech products.
Roy Morejon: Absolutely. Well, Edward, this has been awesome. This is your opportunity to give our audience your pitch, tell people what you’re all about, where people should go, and why they should check you out.
Edward Zhong: Sure. If you are interested in smart home technology or just beautiful products for your home in general, you can check out our campaign. It’s spelled Ayi, so A-Y-I. Just search up Ayi on Kickstarter and our project should pop up, and you can check us out.
Roy Morejon: Awesome. Well audience, thanks again for tuning in. Make sure to visit for the notes, the transcript, links to the campaign and everything else we talked about today. Of course, thank you to our crowdfunding podcast sponsors, The Gadget Flow and BackerKit. If you like this episode, make sure to leave us a review on iTunes. Edward, thank you so much for being on the show today.
Edward Zhong: Thank you, Roy.
Roy Morejon: Thanks for tuning into another episode of Art of the Kickstart, the show about building a business world and life with crowdfunding. If you’ve enjoyed today’s episode, awesome. Make sure to visit and tell us all about it. There, you’ll additional information about past episodes, our Kickstarter guide to crushing it, and of course, if you loved this episode a lot, leave us a review at It helps more inventors, entrepreneurs, and startups find this show and helps us get better guests to help you build a better business. If you need more hands-on crowdfunding strategy advice, please feel free to request a quote on Inventuspartners.Com. Thanks again for tuning in, and we’ll see you again next week.

Join the discussion

This site uses Akismet to reduce spam. Learn how your comment data is processed.

More from this show

Episode 292
Art of the Kickstart logo

Subscribe to the Art of the Kickstart Newsletter

Sign up now to receive the Art of the Kickstart Crowdfunding checklist and notifications of new posts and interviews

Success! You are now subscribed!

Crowdfunding Checklist

Fill out the form below to subscribe to our weekly newsletter and get a free copy of our checklist for preparing for a successful crowdfunding campaign!

You have Successfully Subscribed!