Explore cutting-edge insights in our SuccessCast series. In this episode, Jan Klawer, CMO and co-founder of 1-2-Taste, speaks about their online platform for food and beverage companies to find, choose, and order ingredients, the development of their AI-powered copilot FIFI, and the importance of digitization in leveraging AI. Learn about the impact of generative AI on the food industry and how 1-2-Taste is using it to enhance customer interactions, streamline internal processes, and support product development.
Podcast Transcript
Intro
Host Gerard Klein Essink: Welcome to this podcast on AI and product innovation. I'm here together with my office mate, Jan Klawer with 1-2-Taste. Jan is with his company like us at Wageningen Campus. And we share the same office. It's my pleasure to welcome you to share your experiences here in this podcast. And particularly because you are one of the leaders in this field within the business that you are doing, but I think you can best explain it yourself, Jan.
Guest Jan Klawer | 1-2-Taste: Sure, sure. So welcome for, thank you for having me on the podcast. Indeed, my name is Jan Klawer and I'm CMO and co -founder of 1-2-Taste. Originally, I have a background in digital marketing and data consultancy, but now since more than five years actually working for 1-2-Taste. 1-2-Taste is the world's first online platform where food and beverage companies can find, choose and order all the ingredients they need. So we sell online ingredients like flavors, colors, sweeteners, nutraceuticals, proteins and much more.
Host Gerard Klein Essink: Can you maybe explain a little bit why you started with this company? What was your idea behind it?
Guest Jan Klawer | 1-2-Taste: Well, that's quite a long story. Yeah, for sure. The main reason behind it is to give small and medium companies...
Host Gerard Klein Essink: Long stories can be very interesting.
Guest Jan Klawer | 1-2-Taste: In the food business actually the same access to ingredients as the large manufacturers as ingredients. So typically, in this SME segment it is super hard to find ingredients and to order them also in the right quantities and then that's the basis of it. Ingredients manufacturers actually like to sell ingredients, but they simply can't serve this segment because their processes are not set up for it. So the solution for this problem from both sides is actually digital technology and having a digital platform that brings together supply and demand of food ingredients just as easy as ordering ingredients on or ordering other items on Amazon.
Host Gerard Klein Essink: Yeah, so this morning I overheard a discussion in the office between you and your colleague. You just had a webinar on salt and the customer who was on the phone said that he was very, very satisfied with how you did it, how you managed it. And of course, I also shared the information. So I found it very interesting to hear that and also to have your colleague actually still explaining technical details behind it because many companies in that space still need a bit of support every now and then.
Guest Jan Klawer | 1-2-Taste: Yeah, that's the tech we bring to support development of new products in the food industry. But we have the human aspect as well, right? So, we have really food technology experts with lengthy experience supporting the projects. We're trying to, we call that hybrid and we're trying to best put the digital world and the traditional world of interacting with somebody together.
Host Gerard Klein Essink: So, can you tell, and you work with top class in companies like the one that was doing this webinar. How many suppliers do you have on your channel?
Guest Jan Klawer | 1-2-Taste: At this moment there's 100 suppliers, so 100 companies that actually produce ingredients. And all in all, these are 3000 ingredients that we have on the platform in total. So, 100 suppliers, 3000 ingredients for how many applications? Probably 200? No, I think that's...I don't even know. Yeah, so the major ones are of course the bakery industry, beverage industry, everything we call health and nutrition. But you have a lot of sub applications in that. Of course, there's also the meat processing in the dairy industry and then dairy includes even things like ice cream. So, there's many, many application areas. Yeah. We don't count those.
Host Gerard Klein Essink: Yeah, well, that brings the complexity of it. No, no, no, but I think that's the core of, I think, your opportunity also with the AI solution, right? To help people to navigate. But you're going to explain that yourself in much better language and words than I can do. But maybe the key question I wanted to ask you is, well, you know that there are many opportunities and challenges, right, in AI. So how do you look at it, how AI can help food companies or maybe the companies in your segment?
Guest Jan Klawer | 1-2-Taste: Yes, I think AI is now a quite a logical step in the evolution that many companies went through. So in many industries, if you're talking about food, of course, we went, if you look at decades back, we went to all kinds of processes to make food production easier, just like production in other areas has been made easier. Then there was an automation wave, right? I mean, how can we automate things, not only in the food industry, but also outside of that? So now the use of AI is the kind of the next step in that. And AI originally invented by Alan Turing that is famous for his Turing test just after the second world war. So since then, AI technology has been growing. Why has it been so quick in adoption in the last years actually is because of the growth of data.
And now we have with generative AI, which is the buzz of the last one and a half years, we have a very disruptive application of AI. It's also very easy to use for many people. And I think most people that listen to this podcast have actually used it. So 1-2 Taste is digitizing how we see the supply chain of ingredients to make it much easier to find and order the ingredients you need. So, using generative AI, but also AI in general, is a very logical step for that.
Host Gerard Klein Essink: So, you're seeing this as a next step in the evolution of the food industry and an easy one. I think that is important to understand where maybe ERP systems in the past weren't that easy or convenient to do.
Guest Jan Klawer | 1-2-Taste : Yes. I wouldn't call it easy to completely embrace AI, but to take the first steps with generative AI. I mean, everybody knows, of course, chat GPT by now. It's quite easy to actually start. It's not easy to do the whole process.
Host Gerard Klein Essink: Yeah. So tell me then about your system called Fifi.
Guest Jan Klawer | 1-2-Taste: Yeah, it's nice that you ask about that because we love talking about that. It's our step into the Gen .AI world. So as a digital platform, you already asked, we have thousands of ingredients on the one to taste platform. And we have developed our own AI powered co -pilot as we call it. It's called FIFI, where it stands for food ingredients found instantly. So, it's a very nice acronym we all found. It is built by us. So, our technology development team has built it. And one of our co -founders, he put a lot of time and effort into that. It's technically using the open AI large language model, but it only is on top of the 1-2-Taste ingredients and documentation database. So that's something that becomes very visible externally to customers and visitors on our website. Because if you go to want2taste .com, you'll find our chats, like many other companies have chats.
But if you ask a question for a recommendation for ingredients for a certain application, for example, let's say you want to reduce salt in bakery and you want to have, and you're asking the question there, so I have these baked goods, I want to reduce salt, what do you recommend? It's actually this mechanism behind FIFI that generates the answer which is quite a step from where we were before. The good thing about it is that we can also externally can be used, everybody can use it for themselves, but it can also be used internally by us. So as our sales and customer success teams can actually not only search on our platform, because nobody can remember the thousands of ingredients that we have, but they actually ask internally, Fifi, okay, I have this application, I need to have this ingredient because a customer is asking it. And they can find it much easier than themselves. And I'm, for example, in plant based proteins, we have maybe 60 different products. And I'm convinced that moving from a simple search to actually moving to Fifi, it's much better for the sales team to actually give recommendations. And one nice thing as well that we recently had an example of is that our sales teams are preparing themselves for customer calls because they say, okay, the customer is a beverage company in this space, this size. They like to reduce the sugar in a certain soft drink. And then actually they get that kind of a not a script, but they get the right points to talk about to the customer already by asking it to Fifi.
Host Gerard Klein Essink: Yeah, so in the past you needed to have a lot of experience, right? Maybe 10 years over years of experience. You can now use the database behind it to see what could be a good advice.
Guest Jan Klawer | 1-2-Taste: Yeah, I wouldn't say it's not. So what you then would think is that does it replace a person? But it doesn't. But the typical example of, you know, salesperson hangs up the phone and, I forgot to talk about that and that and that ingredient. There will be a lot less because all the relevant ingredients will actually be shown to the salesperson to already talk about.
Host Gerard Klein Essink: Yeah, but I would think that this is a major benefit for SMEs that they can get some kind of direction for the problems that they are facing instead of having to talk to a very big distributor that might not have the time or the principle itself is not on the phone as well. So, with this they save time and they get a better information.
Guest Jan Klawer | 1-2-Taste: yeah, indeed.
Host Gerard Klein Essink: Yeah, well then maybe the question is, yeah, well, I highlighted it already in maybe my answer, but why is it indeed so successful for you and for your customers?
Guest Jan Klawer | 1-2-Taste: Yeah, from what I already explained, it's much easier to find ingredients for the right application. So if the customer is going into a chat, if you compare it to previously searching, you searched for plant -based protein and you got a list of 60 results, right? Now you're saying I need plant -based protein because I want to partially replace meat in this and this and this application. Then you get tailor -made results, which is probably a shorter list and you don't only get a list of ingredients, but you also get the relevant background information, right? With the key attributes already, links to documentation, links to the products, you use all of that. Just like...You know, searching in Google used to give you a long list of search results.
Google themselves is now moving to what is called a search generative experience to actually have a kind of a summary already at the top. And you even have, maybe some of the audience are familiar with it, you even have tools like Perplexity that is giving you a summary with links on everything that you ask. It's kind of the between searching and chat GPT, it's kind of an amalgamation of these two. So we're trying to bring that as well on our website to the customers. An obvious one, of course, for internally for us is time savings, because they don't need to search around filter and actually when they want to need to do suggestions to customers. It's a lot easier. And at the same time, we also started using Fifi to generate emails to our customers. So that actually a salesperson at the push of a button can say, okay, I want these ingredients.
It is a R &D person in a company that is producing these baked goods and they're in Germany. So actually, out of that comes the base, a very strong basis I would say, of an email that they can actually send to the customer. So needless to say that, well first of all the relevancy of these emails goes up because you can really tailor made it towards what the customer needs but also it is a time saver because you don't need to type that entire email anymore. So on both ends, it is a win.
Host Gerard Klein Essink: Well, that's clear, Jan. I think I've got another question, but maybe I'll save it as the last question. But for what I wanted to ask you, and I'm very grateful for your generosity to share some of your experiences with other organizations. The question is related to this. You know, what are the things that you think are important when you want to get AI to fly in your own organization from what you've experienced yourself? Could you mention a few things?
Guest Jan Klawer | 1-2-Taste: Yeah, I think there's a couple of things I can mention there. I think the first thing is digitization, because one thing we are digitizing the supply chain of food ingredients as one to taste. And what we see with all companies, whether it is customers or is our suppliers, is that digitization is very often very rudimentary. So if you want to go to using AI or you want to use to go to technology actually at all, you need to digitize. I would start with that. Think of your core processes in your company, right? What kind of systems do you use? Do you actually have systems for things or are you still writing things down on paper? And I think a lot of people think, I actually do. You know, how do you log things?
For example, for your customers, do you have a CRM? If you don't know what a CRM is, you probably don't have one. And this is really the starting point. Why you need that? Because if you wanna go into AI, you need data as well. You need data points in a digital way. So, are you actually going looking high level, are you generating that data in your company? Because whatever you want to do in AI, it needs to be fed with data. If you're talking data, how and where do you store all the data about your products, right? Your customers, your stakeholders, what do you log and measure around all your company processes? In what systems is all of these data? Are these systems talking to each other or is it separate spreadsheets somewhere left and right in different departments? Is it all accessible? Well,
And then the third big thing is people. So the training of people, but also the mindset of people. AI and also generative AI is not an enemy, because there's the famous quote, you will not be replaced by AI, but you will be replaced by a person using AI. Certainly, Gen .ai, so using GPT and all these applications around it, is a super simple thing to start using. And I think many people have already tried it. If I ask every time, like, how often do you use it, a lot of people say only once a month and not even once a week, which is very strange. Because whether it's used for even text generation, processing, summarizing, you can give orders to a Gen .ai GPT. And it's a huge time saver that I actually used in the preparation for this interview as well. So I wanted to know what's the length of what we're going to talk about it very easily, Chat GPT is telling me that, right? There's just thousands and thousands of use cases. And everybody can use that. But you can also think of using Gen .ai for image generation or data processing, right? I mean, a lot of things we were doing in spreadsheets. It's a lot easier if you feed it into Chat GPT, the spreadsheet, and tell you, OK, what do you make out of these data? So one of the things we did internally in the company is.
And it's a first simple step is to train the teams about prompt writing training. So a lot of the Gen .ai tools now, they start with giving the right commands, a prompt to do just anything you want. And actually getting all the different departments in your company up to speed and getting them to know, okay, this is how I structure a prompt.
Host Gerard Klein Essink: Simple step is to train.
Guest Jan Klawer | 1-2-Taste: is something that is very, very helpful in their day -to -day work. So as said, we've done this internally and we found it a super useful thing.
Host Gerard Klein Essink: very logical, but this makes me think actually about the times that Google came to us. You were also learning how to search. And this is then of course writing a prompt with AI and having some similarity or some understanding in the company how to do that is very important. So thank you for that. I wanted to come to maybe the last question I had. Now you are here, right? The future is there or the future is there or in another direction. So what do you think AI will be in a few years time being applied to your company?
Guest Jan Klawer | 1-2-Taste: Because you talk about the future, it's definitely, I think, the future is bright. I think there's more positives to negatives and I think this will go into the positive direction.
Host Gerard Klein Essink: you talk about this. Yeah.
Guest Jan Klawer | 1-2-Taste: Yeah, so in 1-2- Taste, we have a couple of things we're working on right now. So from what you've seen, from what we talked about, is we already started with it. We´ve already started with doing it one and a half years ago with the first applications of Gen .ai and the first steps towards FIFI. What we will now do is really embedding AI in our processes. A lot of companies do that. They send, for example, marketing communications out on a regular basis. We have, for example, marketing themes and we had, for example, the sodium reduction, which is a very important topic as a topic right now.
But if somebody that doesn't know us comes to us next week after we finish the sodium reduction theme, all the communications and all the information about sodium reduction that we typically send in calendarized emails, that will not come to this person. So we will use AI in future to actually get the right types of information around ingredients for a certain theme in front of the right person when that person is open to it. Because we might have a calendar in our company, but that calendar is of course much completely different than the calendar of people that are our customers. This is a first step towards it and I think there's many more steps over the next years to come.
Host Gerard Klein Essink: This is the first step. I can imagine, yeah, so you're creating actually a dynamic situation and in a certain way also a long tool for the information that you have already produced or gathered.
Guest Jan Klawer | 1-2-Taste: Yeah, it's dynamic and it's individualized, personalized towards the person. As a marketeer, you think very often in segmentation. There's the famous buzzword of, yeah, the best segment is a segment of one, which is basically individualizing and not all the bakers together and all the dairy manufacturers together, because everybody has different ideas, different priorities, different challenges because otherwise you're only talking about the common denominators and AI will help us in doing that in approaching and solving the problem that that particular company is dealing with around whether it is sodium reduction or clean label or or sugar reduction for example.
Host Gerard Klein Essink: Maybe a question which I'm not completely sure if it's already available in your tool, but something that I know that larger companies are working on. And the situation is like, okay, if you are a larger company, you get all sorts of new suppliers coming to you. They promise you the moon. Well, you are looking for a certain ingredient that will gel at a temperature with a certain viscosity and maybe, well, give a certain color as well.
So do you think in the future that if I would be an SME company, I would come to your platform? I could say, well, I'm looking for my beverage with a pH of this and a maybe translucency of that, a viscosity in this direction. What would be the best ingredient that you could advise me?
Guest Jan Klawer | 1-2-Taste: Yeah, so this is something that right now, FIFI can recommend already. So, if you're a small manufacturer, a medium -sized manufacturer, you go to the traditional distributors and you would ask these questions. I have these challenges with any viscosity.
Well, first of all, it starts with: will the distributor actually talk to you? Because if it's a big distributor and you're called Carg, you're called Mondeles or Nestle, they probably will want to talk to you. But if you're not called Nestle or Mondeles, will they want to speak to you because the first thing they ask is like, okay, how much are you going to buy from us on a yearly basis? But we don't ask that question on our website, right? Everything is already on the website now and fully transparent, including the prices. And then now already you have the layer of FIFI on top of that that can tell you, okay, there's these and these and these ingredients options. If you're talking about what we call technically attributes like viscosity of the end application. What will this product do? FIFI can already advise you in that. But of course, with many more data points and with many more inputs and experiences that we actually have with certain ingredients, we are feeding Fifi with much more data. I can imagine that at some point, Fifi can really advise you as well in product formulations, to a quite a far extent. It's not something we will have next week, but it will evolve into that.
Host Gerard Klein Essink: I think that's a great advice. Well, I think that is.
Guest Jan Klawer | 1-2-Taste: And it will never replace, but it's more like taking the different points and taking experiences. As long as we have the data somewhere in our database, Fifi can actually do it. And with growing, growing data in our platform documentation, experience, product sheets, FIFI can amalgamate that into an advice that will actually help people in the lab like, I didn't think of this, well, let's do that. And then there's always, of course, the human people to talk to. And I think this will be a super nice compliment.
Host Gerard Klein Essink: That's right. Well, with that, I think I would like to thank you so much for sharing all your very valuable know -how and give an insight about where you are, where you're going to go, how you're going to fly yourself as a company, because I think you're already at a very high altitude with your company in terms of the miles that you have made by implementing this and setting up the company. You will also be part of the master class, right? You can face to face explain a few things related to the summit, the third until the fifth of September, which is really valuable and much appreciated. And I look forward to having you there and to be part of the team.
Guest Jan Klawer | 1-2-Taste: Great. Yeah, we're looking forward as well. I just want to taste as a partner. So happy to do the master class. And I hope to meet a lot of people there and to discuss with a lot of people into what AI and food tech will actually bring for them. So, thank you for having me.
Comments