Tastewise in 20 min: Demo of the Tastewise Gen AI platform

Speaker: Amit Ashkenazi, VP Sales, Tastewise

Tastewise’s VP of Sales, Amit Ashkenazi, takes us through the context of data and research in the food and beverage industry and the importance of AI adoption. He then shares a flow of concept-to-product development on the Tastewise platform to illustrate how AI can be quickly adopted into existing workflows within food and beverage research and marketing.

[00:00:29]

Hello everybody, thank you for joining us and welcome to our Generative AI Food Pioneers Summit. This is our second day, super happy to have you. The session of today, we’re talking about AI and the evolution of AI in the food and beverage space. It’s going to be amazing. So let’s dive in.

A little bit about the topics we’re going to cover today. We’re going to start with the AI revolution, what’s been going on. We all heard about it. We all heard about the recent tools that have been coming around in our workspace. Let’s talk a little bit about the 

[00:01:00]

impact that has on our lives. Then we’re going to take a focus on the food and beverage space. That’s, of course, what we’re here to talk about, so taking that in the food and beverage space and then a really exciting workshop, if you will. Can AI and generative AI be put into use to go from concepts from a high-level trend even or a vague concept all the way to a product, all the way to a finalized product? Can AI help us go through that journey?

This is going to be the gist of our sessions today to make sure we put this to good use and we see how that can actually be impacting our lives. In terms of introducing a little bit of our clients, our partners that are all trusting us and are working with us to make sure they adopt these new technologies and are basically a part of this revolution.

But as I said, I do want to start a little bit of a one step back. One step back, I mean, the revolution that happened to our lives in the data world even 

[00:02:00]

beyond the food and beverage data space. I like opening up with this, again, quote from Forbes, 90% of the data in the world has been generated over the last two years, in two years alone. So again, 90% of the data has been generated over the last two years. This is, when you come to think of it, a really incredible statistic. But when we talk about tracking trends, the speed of trends, all the big data trend that we’re all a part of, 90% of that was also only in the last two years. And that speaks very strongly into the speed in which things develop today. The speed, the exponential speed in which new technologies are being introduced, new platforms are being introduced, and how fast are these being adopted by our industry.

I think as the hero example that we all think about comes our beloved friend, ChatGPT. So if I were to ask ourselves, how long did it take ChatGPT to get to a million users? The answer is actually five days. So within five days this amazing tool has been getting adoption of a million users. And not less incredible than that, the time it took to reach 100 million users. 

[00:03:00]

So approximately a third of the US, that was done in two months, in two months alone. Let’s look at these amazing logos that we see. We can see Google Translate, we can see Uber, Spotify, TikTok. This is not that these companies are not as successful as ChatGPT necessarily, but it shows about the speed in which things are adopting today, adopted by consumers that we’ll talk about in a second, a little bit of a spoiler, but also, of course, adopted by the industry as a technology.

Let’s take a quick brief on a few examples of how these lives are being revolutionized, especially by maybe younger, non-younger generations. So we almost have to mention Fortnite here, right? Fortnite, of course, as a game that actually revolutionized. 

[00:04:00]

This is actually an event that happened in Fortnite that I think is a beautiful example here. This is what they call a live concert, but of course, a virtual concert, a virtual concert by a performer. Until today, it was one of the most profitable events that ever happened in Fortnite. And people were attending the concert. “The event was years ahead of its time.” I specifically like this quote, “I’ll never forget the hype and chills I got when I attended this concert.” Now, I’m thinking to myself, dude, you were in your room. What do you mean you got the hype and chills? It’s not — you went to a Taylor Swift concert, fine. Swiftie here all around, so we can do that. But I will say, from their perspective, this is their experience. They were in their room, in their virtual space, but still got the hype and chills. And then the only question we’ll have to wonder really is how much are we there?

[00:05:00]

So the consumers are there. Gen Z-ers are playing countless hours each week of computer games. That’s what they get their influence from. Are we there? Are we meeting consumers where they are? They are playing that. Are we speaking to them in the language they want to understand?

Another incredible example, this example from Morgan Freeman. “I am not Morgan Freeman. And what you see is not real. Well, at least in contemporary terms, it is not. What if I were to tell you that I am not even a human being? Would you believe me? What is your perception of reality? Is it the ability to capture, process, and make sense of the information our senses receive? If you can see, hear, taste, or smell something, does that make it real? Or is it simply the ability to feel? I would like to welcome you to the era of synthetic reality. 

[00:06:00]

Now, what do you see?” So, of course, this wasn’t Morgan Freeman. This wasn’t Morgan Freeman. It was AI in Morgan Freeman. As a great Morgan Freeman fan, it was incredible. And the examples of these are going on and on. There are so many other examples of how much technology has evolved. It really left us wondering, again, how are we going to adopt it? So if you think about ChatGPT, ChatGPT is actually only in the bottom left here of the text. We have Bard, we have Copy.ai, we have a lot of other tools in the text. The industry is really being revolutionized by generative AI across many, many other terms, whether it’s creating videos, voice generation, the deep element, the deep fake, what they call, but the deep translations that exist, images, coding. 

[00:07:00]

All across the industry, we’ve seen more and more industries are adopting this technology. And this is why I don’t think it’s actually surprising that the food and beverage is no different. The food and beverage industry has been revolutionized by AI as well. And actually, this is where Tastewise comes in. What Tastewise has built over the last five years is a generative AI system that makes us understand consumers. So let’s see how this is being translated into the world today.

What I want to speak about before we dive into that is actually about the consumer of the food and beverage space. Because, as I said, the industry has evolved, technology has evolved, but it’s actually only one manifestation of what we see in the industry. So three parameters or three categories of influence that we see or we see has been revolutionized of the consumers itself in the food and beverage space. One of them, the first is about influencers. How do we actually decide what we eat to think about. How do we decide? And we all know that we have higher expectations of our food than ever before. 

[00:08:00]

It’s not only to make us to make us full. So if in the 2000, it can be a cupcake in Sex and the City of Carrie Bradshaw, as we continue to the world of influencers, that everybody’s becoming an influencer. We have actually Instagram pages today that are actually AI. They are the machine learnings. They’re not necessarily real people. And we see that so typical. You see that picture to the left where here’s a huge selfie, the guy with the long hand, the person with the long hand that actually can take a picture of their food and share it with their group. So that’s about the influencer side. The actual assortment is a critical piece. The milk, which was traditional, became a retail shelf with 20 or 30 types of milks. The way to get it can actually become delivery platforms. So it’s not only I need to go to the supermarket. Yeah, we can order. Everything can be instant. And of course, the consumers, as we said, they no longer accept health. They expect health. They expect gut health or brain function. How many of us have heard during COVID the mental health rise? 

[00:09:00]

And we’re left to wonder which types of food are connected to which types of mental health claims. But the reality is they expect so specific benefits from the industry to achieve, which is why tracking what they do maybe becomes harder than before. Harder than ever, I might say. One piece is about the richness. The other piece, as we said, is about the speed. So a quick snapshot of maybe some trends that we all heard, we’ll continue to hear about. But look at this amazing curve of all of the trends that have basically developed over the past decade or so. Not a lot of them had a lot of lead time before that. So how long did it take CBD to become CBD? When the TikTok feta cheese recipe, I’m not sure how many of you guys remember that, when it went viral, it was COVID times, we’re at our homes cooking. All of a sudden there was a shortage of feta cheese in supermarkets and retails. So these trends are evolving, showing one, the specificity,

[00:10:00] 

the high expectation of the consumer, but also the speed in which they develop. So this is what the AI is actually allowing us to do, to move at speed and at scale of any activity you want to wish to do in the organization. The consumers tell us what they want to do. We just need to be in the right place. And on that end, I want to separate between two different types of inputs, of insights that are actually used in the industry. The first one, I will say even more traditional, is what we call shopper insights. 

You’re walking into a retailer, there’s so much input going into what is actually happening there. First of all, how much we’re selling, how much the category sales is growing. All of us use that information in most of the presentations we make. Are the dollar sales of our category increasing, decreasing? When do we buy the product, summer, time of the week, whatever? All the way through, how many dollars should I add to my promotion to maximize the share of wallet of consumers,

[00:11:00] 

to maximize their spend before a holiday? That’s I think one side of the picture. I want to ask, what happens when you go into your kitchen? You check out of the retailer with these three products. What actually did you cook? This is where we complete the picture, if you will. No pun intended. We go to the consumption moments and we can see that the consumers that are cooking from these, like the three dishes that can be created, are just very, very different and have very different ramifications of why they choose them. Because if I’m cooking a healthy chicken taco, quinoa salad or mac and cheese, the reasons I’m doing that are very, very, very different, which comes the why, comes the context of consumers, which we have to understand. An amazing way to understand that will be to look at what people do. Instead of asking them a question, look at what they post on social. Look at the best selling menu item on delivery. Look at the recipe itself. What exactly are they doing? In this example, we have a user named Jenny. She’s from Phoenix. She’s a mom. She’s telling the world about the dinner she cooked for her kid. 

[00:12:00] 

My four-year-old loved the vegan creamy kale cauliflower gnocchi. Now, imagine the next time you need to know what do people eat with pasta. Imagine somebody reading 100 million of these and telling you what’s texture, what diet, what vegetable they were actually including, all the way through connecting it with the recipe and the food service dish that was actually presented.

Enough with introductions, let’s dive into what we actually mentioned in terms of the concept of product. We want to take a fun concept of a yuzu cream cheese. Again, yuzu is maybe the new kid in the block again that’s rising from the families of the superfoods, an amazing citrus here. How can AI help us create a full workshop from a concept to product? Before I dive in, one of the best I think examples we had was a session similar to this that we did in a framework with catalogs. And the main value there was they were telling us this was taking us three months before. Three months before to understand what’s the concept, why, who, what’s the recipe we can actually do. This goes into three days. 

[00:13:00] 

Now let’s actually do it in 10 minutes, which is even more challenging. But, again, let’s see how the technology can be put into an amazing good use in the day of the life.

Let’s start our innovation process. We start the innovation process. We actually want to understand which cuisines are rising in terms of consumers and what they do when they cook with cheese. Very easily I’m able to understand Japanese here. It’s not the most popular one, which is great, but it’s the one that’s rising the most. I already have an essence of what cuisine I maybe want to focus on. Then let’s start drilling down, which is going to be something we’re going to do a lot throughout the next few minutes or this storytelling. What ingredient is actually being dominant when they cook with Japanese and cheese? So they cook with — we know well on cheese. We know Japanese cuisine is dominating. What are the ingredients that are being used? Here comes the yuzu. The fruit and vegetable here, popular than the rest, still very much growing, 

[00:14:00] 

basically validating my assumption that I want to continue to do that, continue to invest in that concept. And, of course, I’m establishing the understanding that cream cheese as a category is something that will greatly fit with consumers. Why? Because they already say it. And when I say they already say it, they already do it, I don’t necessarily need to go out and check the evidences. All I can do is look at the examples. This is actually a real gift from our platform. When we do yuzu and cream cheese, these are actually the examples. Not only these are the examples, we can deep dive into the most popular consumption moments, popular moment. This is actually a bestseller. So there’s already a dish, which is the butter toasted cake using yuzu marmalade cream cheese that’s a bestseller on Uber Eats or DoorDash or Grabhub. So again, very easily, instead of maybe guessing or — just a second, amazing. So we can look into these examples to see what people do already. And people do amazing things. 

[00:15:00] 

We all know that. They’re creative. They’re expecting more and they’re creating on it. Let’s look at these examples from, again, social or recipes or menu items of what is actually happening to make us more knowledgeable and meeting our consumers again where they are.

Continuing with our story, we have a lot of inputs. We know yuzu cream cheese is validated. We have looked at the most popular examples or food service or the actual social posts. How do I deep dive and answer a lot of the related questions? Here we actually do what we call an AI survey. So it’s a regular survey. You need to go and you ask 500, 600, 1,000 people, hey, would you want to eat this? What diet were you following when? The reported behavior, how can we actually use AI to do observed behavior? So imagine having a survey of people and what they actually did. And with that, you can actually survey 160,000 people. That’s what we actually did to educate ourselves about yuzu cream cheese. These are all the questions that actually came from our data. Which of the following 

[00:16:00] 

best describes or in which category would you like to see yuzu? Again, the input might come from people that actually do it. Where are they already inputting yuzu as a new ingredient? Continues on how important are the following ingredients or motivations for you? All of these people already speak about. So an amazing way to understand is to deep dive and to enrich our knowledge about the concept we want to create is actually not running a survey, running an AI survey that will come back to us with answers almost instantly.

Let’s continue with our journey. Man, this is exciting. Continue with our journey. So we have a learning, we have understanding of the consumption moments. We have ran a survey that actually allowed us to understand, can I now continue and let the AI create concepts for me? Say it again. Let the AI create concepts. Can AI do that? Look what happens when you type in yuzu cream cheese in our platform. This is what we call the 

[00:17:00] 

Concept Discovery. Three, actually of course there was more, but three ideas that can actually be implemented in the market. For the people wondering, the images here are not real. It’s AI or sometimes we say they’re not real yet. But it’s generated by AI live on the spot. Not only does it give the actual concept, it’s giving us a way to open the black box by explaining why. Why was this chosen? So it was chosen because this is the related claim of sour flavor in the one at the bottom that’s actually growing in consumer mindset. They’re already incorporating that. I sometimes say, welcome to the new world of content creation, of concept creation. You don’t like this concept? Regenerate. Regenerate. I’m not saying any company would actually do maybe this as is, but it will support it and maybe replace half of the work or 60% of the work. So we have a concept and as we said, it doesn’t end there. We always have a surrounding, a whole picture we want to paint.

[00:18:00] 

In this example, I want to take to say, how to position this, let’s say, which audience of consumers might resonate with the yuzu cream cheese as a new introduction to the market. Again, can AI help me do this? Here are two suggestions. Here are two suggestions coming from AI about how can you position your cream cheese? One, for adventurists, for Gen Xs combining with roasted. Again, you always understand why. Or as a gourmet product for the foodies amongst themselves, about a gourmet experience because it’s combining the yuzu cream cheese. So we have our concept. We have our positioning. How would people use it? What are the demand opportunities? Let’s dive in. Yuzu cream cheese with scallions, yuzu cream cheese with radish dip. This is where we tell the AI, I have my concept, I know my positioning. How do I incorporate that? What resonates with consumers and what types of — oh, my God, this is a beautiful one, the spinach croissant. 

[00:19:00] 

So you can see it’s bringing in a lot of different inputs. It can be from the sweet world category, from the savory category. It’s all based on what people already do and what will resonate. But it giving us, again, that richness. It’s not a product as a standalone. When I talk about it, when I speak about it, I know who am I addressing, like the audience. And I also know how can I recommend people using it. And again, I don’t know, infinite amount of applications to that.

So we have our concept. I’m going to start actually bringing it to life in terms of I want to cook it. I want to go into my kitchen and give my chef the ability to cook it. Now, they can go in and try to look how to combine yuzu with cream cheese or can the AI give them a recipe, which is very, very amazing? This is a recipe, yuzu cream cheese tart with honey drizzle that the AI came up with. Again, it’s about the recipe to shorten

 your work at the speed in which it will take you to do that. But again, always anchoring ourselves with what consumers want to see and what resonates with them. Here is a reaffirming of what we started with. We have the 

[00:20:00] 

Japanese point that’s growing more and more in consumers and the more applications they do. Yuzu and cream cheese and honey, all ingredients that are continuing to grow, which is why the recipe has added them to the project, to this recipe specifically.

I just came back from a chef conference with Kraft Heinz, an amazing event. We cooked the sauces of 2027 and beyond. Some of the inputs were recipes generated by AI. The industry is really adopting these elements. And the chefs were the first ones to say, wait, how? I need to change this, change that. Incredible. So we have a recipe. We want to go into the execution phase. Let’s write a blog. Let’s write a blog. Maybe this is the closer one to ChatGPT that we know but again, only on what consumers really eat and the Tastewise data, of course. I want to go to the social. I want to have social posts 

[00:21:00] 

and promote myself about yuzu cream cheese. I can reference the examples I saw before from the recipes and social posts. I can actually generate new content. Again, of course, based on what people connect, whether it’s Mediterranean diet, lactose-free, whatever that may be. I can already use the AI to get to the final execution piece. I think the end all be all, can I actually get my product in?

And this is the — we promised a session from a concept to product. So we started with a concept, here’s our product. Here’s a yuzu cream cheese branded for great value of the Walmart private label, but, of course, can be done themselves. More than the actual picture, think about the amount of questions we usually have when we want to develop that pack. We want to know what is the description, of course. What do we actually call out on the pack? There are probably 20, 30, 40 different options of what to call out. Calorie value, the diet that they want to follow, what should I highlight in the one-liner? What if you can have a starting point? You can have an edge by, again, giving the AI 20, 30 inputs that you’re considering. Let it help you prioritize based on what people did, based on what people resonated with already in 

[00:22:00] 

similar scenarios. This is what we mean by revolutionizing the industry. Every decision we can make can probably be made faster, for sure, and definitely at scale as we continue to grow with this.

As a quick recap before I go into my very last application, we started with the concept of understanding what do people even eat with cheese? We’ve discovered the Japanese element as a cuisine. We saw yuzu as an ingredient. We saw the connection to cream cheese. We’ve created concepts. We understood what is the audience that I need to go to. How will they use it? Is it a savory or actually recommendations of the radish dip? We created a recipe, created the blog, created the pack, all within, whatever, 10 minutes or so. And, of course, a full workshop can be really done live. The last ending of this is actually how can AI support me in the distribution? Or when I go to talk to operators and I want to pitch this new idea, provide customized recommendations. So let’s say you take this product and you want to

[00:23:00] 

sell it to the Cheesecake Factory. Makes a lot of sense. You want to come up with a thought leader content for Cheesecake Factory specifically. This tool of what we call the menu item generator combine the best of both worlds by, it’s reading the menu of your operator, in this case, Cheesecake Factory, and it’s offering dishes specifically to them based on what they have on their menu. It’s a big, big, big change. So, for example, if this is offering chicken enchilada, it means that it’s already on the operator’s menu. So you’re coming in with your product that you want to pitch with all of the storytelling you have done, but you’re not even showing things that will require them to go outside and change a lot. All you need to do is actually add this. The rest are already in your kitchen. Talk about the ease of implementation, the readiness and the easiness to digest that information easier for any operator that they don’t need to start ordering new ingredients off. So you are basically making your recommendation more 

[00:24:00] sticky and, of course, more customized all in a click of a button.

It’s a new world, but it’s always challenging ourselves. What does it help me? What can I do better? How can I meet my consumers better or meet them faster and make all of my processes from a high-level concept, from a trend that I want to follow, all the way to the actual recipe blog product that I want to be able to create?

Hope this was valuable for you guys. We’re actually ending with an amazing gift. This QR code is going to lead you to our free trial, so you can actually test it out. All of these generative AI tools that I created today that I showed you guys today, typing in yuzu cream cheese and develop your concept, understand your audience, how do consumers eat it, develop your recipe, feel free to go in. It doesn’t really require to be an expertise. I’m definitely not a chef by any means or very far from it, but

[00:25:00] 

give it a shot. Give it a shot and feel free to let us know what you think. Thank you so much for being with us today. It was a pleasure and let’s have a great rest of our summit. We’ll talk soon.