There is a massive amount of waste in the fresh produce industry - of all fruit/veg grown, only 55% makes it to the store shelves. The other 45% is wasted. This is the problem ClariFruit is aiming to solve using a combination of computer vision, AI, and analytics.
To find out more I invited Elad Mardix, CEO and co-founder of ClariFruit to come on the podcast to tell me more.
We had a truly fascinating conversation and, as is often the case, I learned loads, I hope you do too...
If you have any comments/suggestions or questions for the podcast - feel free to leave me a voice message over on my SpeakPipe page or just send it to me as a direct message on Twitter/LinkedIn. Audio messages will get played (unless you specifically ask me not to).
If you want to learn more about how to juggle sustainability and efficiency mandates while recovering from pandemic-induced disruptions, meeting growth targets, and preparing for an uncertain future, check out our Oxford Economics research report here.
And if you want to know more about any of SAP's Digital Supply Chain solutions, head on over to www.sap.com/digitalsupplychain, and if you liked this show, please don't forget to rate and/or review it. It makes a big difference to help new people discover it. Thanks.
And remember, stay healthy, stay safe, stay sane!
Essentially introducing to this industry next generation technologies such as AI, big data, and computer vision technologies that's going to help the stakeholders make decisions based on data, rather than based on intuition or experience, which is actually what's taking place today.Tom Raftery:
Good morning, good afternoon or good evening wherever you are in the world. This is the digital supply chain podcast, the number one podcast focusing on the digitization of supply chain. And I'm your host, global Vice President at SAP. Tom Raftery. Hi, everyone. Welcome to the digital supply chain podcast. My name is Tom Raftery with SAP and with me on the show today I have my special guest Elad elide. Would you like to introduce yourself?Elad Mardix:
Sure, thank you, Tom first for for having me and ClariFruit on this platform. We're very excited to be here. I'm Elad Mardix. I'm the CEO of ClariFruit. We've founded ClariFruit about three and a half years ago with two of my other partners, before ClariFruit I actually spent about 12 years in a very different industry, which is investment banking. I used to work for JP Morgan, in technology, investment banking, about eight years in us. And in the last five years, I build and ran. JP Morgan's technology investment banking business in Asia Pacific. In 2017, retired him came back to Israel to join my two co founders and establishedTom Raftery:
that's an interesting transition from investment banking into Clara fruit. How did that happen?Elad Mardix:
That's a good question that a lot of people asking me about, especially if it is very unique. So as an investment banker or technology investment banker, particularly, which is what I did, for 12 years, you actually spend a lot of time both with large companies, as well as with smaller companies. The smaller companies are usually young, very high growth, private technology companies, who are getting close to the public market route about to IPO or about to do a big transaction. And especially in my last role in Asia, I had the privilege to work with honestly some of the most exciting startups that went public, if you think about it, even Alibaba, in 2000, I think 11 or 12, we we took public Well, today's a behemoth. But, you know, 1520 years ago, it was a start up. And I had the privilege of working with a few other companies, which I took public. And I think towards my career, I kind of decided that this is actually the more exciting path for me to try as my next app, rather than stay on the advisory or investment side, and really try to build a business that number one is going to do good for the world. And number two, is going to be exciting to build with some other partners. And that's how I decided to end up on the Purdue intrapreneurial side about three and a half years ago. And honestly, I couldn't be happier with that decision.Tom Raftery:
Fantastic. And for people who are unaware who are not familiar with ClariFruit, can you tell us a little bit about what ClariFruit is and does,Elad Mardix:
of course, so ClariFruit is actually the world first AI powered quality control software platform, what it means, but we effectively build the first quality control software platform for the fresh produce industry. I'd love to talk a little bit about the background of the issue. But essentially what we're trying to do is to help the fresh produce industry, transition from quality control processes, which honestly are still stuck in the 1950s. The entire processes are manual, and there is very limited data that is being generated. And we're essentially introducing to this industry, next generation technologies such as AI, and big data in computer vision technologies that's going to help the stakeholders make decisions based on data rather than based on intuition or experience, which is actually what's taking place today.Tom Raftery:
And why is that important? Okay, soElad Mardix:
let's take a step back and try to understand thirsty industry and any potential issues. So the supply chain of the fresh produce industry is composed of four key players. It starts with the grower of the farmer, who is the guy who's actually developing the crop right growing crop. He then sells it usually to a marketing company and brand. If you think about this sanlucar, the Pink Lady, the door of the world, most people who come out of the industry may think that those are actually the growers of the supplying in most cases do not, they're just the brand who buys the supply the cloth from different growers across the world, they then perform a QC and if they're happy with the quality, they will put the stamp of their brand. That's the marketing company. And the marketing company can essentially sell the supply further down the value chain to either a wholesaler, or retailer, which are the Walmart, Tesco, marks, and Spencer awards. This is essentially the supply chain the four players across it. Now, what's interesting in this industry, in different from most other industries, is that the quality of the supply is the predominant driver of price. You You can think the price is obviously driven by quality. But in the fresh produce industry, What's really unique is that the quality is actually changing on a daily basis, which makes it much more complex to assess the right price for the quality in any one of those steps or the stages in the supply chain. As a result of this issue, the fact that quality is changing on a daily basis. And as a result of the fact that there are actually no standards, no quality standards today that are accepted by all the industry stakeholders, there is a lot of waste taking place in this industry. When I say a lot, I mean about 45% ways from the grower down to the front store of a Walmart. So for every 100 grapes that you or your grower have been growing in their field, only 55 land in the front store for you know, as I told you, I was a technology banker for many years, and I took a lot of big companies public. And honestly, I had never seen such an inefficient supply chain in my life. Before I was introduced to the supply chain. That was actually probably one of the key reasons that captured my interest for eautiful. And so the next question you're asking yourself, when you see those numbers, you say, Well, what are these guys have been doing about it? Right, because this is a lot of waste, it's extremely inefficient, they lose a lot of money. And the stunning fact is that actually, not a lot has been done in the last 100 years. That means that the way the supply chain from the grower down to the retailer, the way they are evaluating quality is very similar to the way they have done it 80 years ago, they use sizings, then they use color palette and there is a human Inspector, who use those devices in order to come up with a view on what's the quality. And then they actually either keep it in a stack of papers in the drawer or best case they upload the information to an Excel or a kind of electronic form similar to excel. And that create a lot of issues which the common denominator essentially it creates quality mismatches. Right so there is a mismatch between the seller expectation of the quality of the supply, and about the buyer. Evaluation of that supply in that create a mismatch and rejections and price adjustment discussions across the entire supply chain. And all of those issues translates to about 7% of financial loss for each one of the players in the supply chain farmer marketing company, wholesaler, and retailer. And by the way, the 7% financial loss is about double the net income margin that these guys are making. And that's that's really the heart of the problem, a lot of inefficiency. And fortunately, no technology has been adopted so far to try and mitigate and resolve those issues. efficiencies. And as a result of that, a lot of produce waste and money is being wasted. And this is essentially where we come in to the game.Tom Raftery:
Okay, so keying that up nicely. When you come in, what is it that you're doing?Elad Mardix:
Okay, so first, we needed to identify what are the core issues that driving these qualities, because we did have the privilege three and a half years ago, to build a platform really from the bottom up, right. And so we identified four key issues. Issue number one was the fact that the actual inspection process is subjective and inconsistent, because the human being is the one who is collecting the data and making the judgment on the quality, that's not ideal, because you told me thing that the quality of a supplier expecting an RFA, and I made expected five minutes after you and I think it's a minus, and that's going to create a rejection, which is going to be very unfortunate for you. So that's Problem number one. Problem number two is that the actual inspection process today is very slow, because it's human based, and he doesn't generate any data, as well. So there is very limited analytics that you can run on the data that's being entered. But probably the most acute problems in the supply chain are is the fact that there is lack of quality standards that are accepted by the entire supply chain. So people speak different language. And we needed to build a platform that's going to address all of that. So what we've done is we build a software platform that's composed of two components. component number one, is a mobile app for the inspector, the mobile app actually utilize our own proprietary computer vision technology that allows the inspector to take a picture throughout the mobile app and a smartphone camera, and analyze automatically the external attributes of the fruit. And obviously, at the end of this inspection process, the inspector is handled with a grain that is calculated by an algorithm. So that's component number one. component number two is what we call the management and analytic dashboard, which is essentially the brainy section of our platform, our back end. And that's where we allow our customers to QC or the operation directors of our customers, to manage the entire process to define the quality standards, to dynamically manage them, and most importantly, to derive real time analytics, that's going to help these guys make decisions in real time. Based on that, in just a couple of examples, with our platform, because the data that is being entered is structured data, rather than unstructured, which is what happening today in the world. The QC director, or operation manager can make decisions in real time, like, for example, who is the best customer to shift the supply chain, right. So they can either avoid a rejection, or they can maximize, I call it maximize the wallet opportunity with the customer. They can also perform vendor evaluation analysis, which are just not available today. Right. So with cloud food, for example, because of the fact that we have the data coming from our own mobile app, and not from a human being, you can actually figure out if you're a retailer, or a wholesaler, who provided to you the sweetest table greets, in the last 12 months, you can figure out who provided you the largest red chairs in the last six months, and so forth, right. And so the the, the availability of potential analytics based on the row data is really unlimited. And then then this is where the Big Data platforms comes into the game and can really transform the way these AI making decisions.Tom Raftery:
Okay, interesting. One immediate question that comes to mind is you gave an example of a quality QC manager being able to look over the last year's data and see for example, which supplier gave the sweetest grapes in the last year, I can understand how you could use computer vision to go for the size of the chair But you're obviously not using computer vision to measure the sweetness of the grapes. How is that done?Elad Mardix:
Okay, that's a good question, because you're right about the observation now, but you're hopefully not going to be right about it three to four years down the road. So let me explain what I mean. So today, you refer essentially to the inspection process the data collection process, in order to evaluate the quality, let's just say of a table rape in the distribution center of a retailer, right. And this inspection process is primarily composed of two components. One is analysis of external attributes could be size could be color, stem, color, etc. The second component is internal attributes, it could be presented sugar or grease as it named in the industry fairness, acidity, pH, a bunch of different attributes. Today, the way we manage it is, for the external attributes, you're going to be able to capitalize on our own computer vision technology and actually complete this analysis literally two seconds with by taking multiple pictures for the internal attributes. At this point of time, the customer still needs to use here, existing scientifical devices. So in the industry today, they're already performing this inspection process. And they use refractometer for weeks on dura mater, for firmness, and etc, etc. What they do today is they use those devices, they measure those attributes, and then the inspector will enter the data either into the Excel or electronically. With cloud fruit, the inspector is still going to use this device in order to measure the attributes. But she's not going to need to enter the data anywhere else, she's just going to touch the device to her smartphone, and all the data will be transmitted wirelessly into the using essentially existing technologies in both devices, whether it's NFC chips that can speak to each other, or Bluetooth protocol. So that's what's happening in the short term. And by the way, because we are a pure play software platform, we are totally hardware agnostic. So we can effectively integrate to any hardware device that a customer is using, and build these wireless transmission capabilities. By the way, we also integrate to what's called new devices, near infrared technology that can for example, measure presented sugar bricks of firmness in a non destructive way. And that's essentially innovation, which is not ours. But we do introduce you to a lot of our customers that don't have the access to that. Now, remember, I told you before that you hopefully going to be wrong if your observation three, four years down the road. And what I meant by that, is we also cooperating with a couple of other hardware providers who work on a technology called hyperspectral. So hyperspectral cameras already exist today. In the industry, by the way, essentially in big soda machines, and hyperspectral cameras can actually analyze through the picture, not only external attributes, like we already doing today, in our app, but also internal attributes, like bricks, and firmness, etc. The issue is that these cameras today are big cameras, they cost about 12 to $15,000 pair camera, but most importantly, they can only operate in a sterile environment, you know, like darkroom, and fixed distance, it's not appropriate, actually, to the way customers performed QC along the supply chain, whether it's in the farm, or at the entrance of a packing house or at a distribution center. For that you want to have the hyperspectral camera in the smartphones. And I do believe that we are going to have that in three to five year timeframe. And when we do have a, you're going to be able to use the cloud food and take a picture based on that technology. And in one picture, you're going to be able to analyze the overall quality of the fruit. That's one derivative. The other derivative, by the way, which is much more interesting is that when the hyperspectral technology will be introduced into smartphones, then you me, our wife, and everyone else is going to be able to go to the supermarket. And before you buy the stable grape, or watermelon or mango, you're going to be able to take a quick picture and figure out Is it sweet enough for you? Is it tasty enough and do I want to spend the dollars buying we're not there yet,Tom Raftery:
but we are on the way Wow, fascinating. And that kind of brings me to my next question, which is, is this in your application? Are you selling it to the grower? Are you selling it to the retailer? Are you selling it to the consumer? You know, who's your customer for this? Or is it all of the above?Elad Mardix:
Okay, so, no, it's not all of the above. Because by the way of the technology requirement that I just mentioned, but So first, let's explain who is performing QC for the quality control, I will divide the the different words that you mentioned to professionals value chain versus not professional, which is the consumer, we do not address the consumer this point, because honestly, we think it's still going to be too complex for the consumer to utilize the cloud for that, which will require the consumer also to use other devices for internal attributes, it's not going to work out it's not easy enough, we'll wait for when we have hyperspectral. However, the professional supply chain, which starts with the grower, or farmer, and goes down to the retailer is already performing this quality control on a sample basis across the entire supply chain. Actually, today, there are about 12 to 15 inspection points that those players are already performing quality check on a sample base. By the way, one question that I always get is about sorting machines. So this sample based QC that we're talking about is separate is in addition to the sorting machines that usually the big packing houses for sample these categories have. So even for an apple producer, right who have a product, the marketing company will have a sewing machine, they will do a sample based QC before the sewing machine, it's called intake. And they will actually do after the sewing machine, it's called pre shipment before they ship it to the retailer or the wholesaler. So in addition to the solid machine 12 to 15 inspections points today, okay, all of those QC is being performed in a manual way, as we discussed. And we're actually going after all of these guys in trying to help them convert their existing manual subjective and inconsistent QC process and make it automatic, objective and consistent. We've started with the big marketing companies, think about all four pineapple, think about Pink Lady for Apple, think about have sex and sanlucar in in Europe for a bunch of different products category. Those are the big brands that people know. That's how we started to actually build our automatic produce portfolio one by one about two and a half years ago. But today, we actually have the ability to digitize the inspection process for all produce categories in the market. Not everything is going to be automated, ie not everything is going to be analyzed flower computer vision technology at this point of time, about 50% of that will. But all the product categories can see on our platform and be defined with a QC process and digital quality spec, etc. And so what happened, interestingly enough, especially through the corona pandemic, is about a year ago, late 2020, after people realized that COVID is here to stay at least longer than they anticipated, is we had a lot of inbound from retailers who figured out they can send their inspectors to their suppliers anymore. And it's very inefficient for them to update those quality standard PDF every week or a month and send it to their suppliers. And so we're actually working today with multiple retailers. Three of them, by the way, are out of the top 20 retailers globally, then we essentially help them manage their fresh produce supply chain and quality through the cloud platform. Summarizing your question we addressed the tire professional supply chain. We're very focused on the retailers wholesalers and the marketing companies are mostly the buyers of the supply. But we are working with a few growers as well and if time goes by, we will expand into the entire supply chain including the growers. And the reason at the end of the day is simple. Our mission as a company is really to help the supply chain make a decision based on data and reduce waste, that's really going to be beneficial for everyone, including us, the consumers, then the ultimate way to reduce the waste across the supply chain is by applying an automatic testing concept platform, like cloud foot across the entire supply chain. So when the grower performing QC, and the retailer performing QC A day after that in their distribution center, they all use the same QC platform, the same computer vision technology and the same quality standards. If that would be the case, I think we'll see a dramatic reduction in waste. And obviously, in improvement in in the cost for fresh bodies for the entire world.Tom Raftery:
Fascinating. I recently recently had Luis Mathias on the podcast from green chain, and they're doing something slightly different, they have a different take on this kind of thing. It's about introducing transparency into the supply chain, particularly for the growers, so that they can show to the buyers are maybe show to the the the the lenders who have lent them money that their crops are, you know, up to a certain quality standard. is is is there a need for that kind of thing? And is that something that you would roll out? Or maybe you have rolled it out already, that the growers can demonstrate to the marketers, for example, that this is the quality of my crop ahead of the marketers actually buying it? That's a terrific question. SoElad Mardix:
the answer is definitely yes. And by the way, transparency is part of our value proposition. But the unique thing with our platform is transparency. Essentially, it's part of the outcome of using cloud first, because talking about them talking about actually how we're doing automatic QC reports in the sec, but the point I was trying to make is that we provide much more than transparency, we actually enable the grower or the marketing company, or the wholesaler, the retailer to perform their QC their inspection process, with a platform that's going to yield a much higher quality inspection process. Because it's not a human driven, these are computer vision or machine learning, machine driven. And so the inspection is going to be objective and consistent. And that actually, by itself, should mitigate a lot of the rejections that we see because it removes essentially the human factor. Now, the transparency factor comes into the play as well, because what happens with cloud fruit is that whenever the inspector completes her inspection, she needs to click done at the end of the inspection process in the mobile app, and the many the second she clicks done. Essentially, there is an automatically generated report QC report that is being wired to whoever actually the user or the QC manager of the user has predefined, it can go to your buyer, it can go to your grower, it can go to the sales people that are in your organization, it can go to the procurement managers in the case of the retailer. And what it allows is couple of things. Number one, if you're a grower, let's take the glory of a sample that you mentioned, you can actually send an automatically generated QC report, which again, I'm really emphasizing is incorporating data that was captured in an objective and consistent way, it's not the grain wasn't dictated by the inspector of the law, it was driven by the algorithm that by the way, the buyer actually came up with right. And it can go to the buyer and essentially say, look, this is what you're gonna get right here is the great by the way, this is the great based on your quality spec, not based on my quality spec, Mr. buyer. And here are all pictures that were taken as part of the analysis because the computer vision needs to have pictures in order to analyze the quality, including defects, pictures, and essentially allow the buyer either to bless it and say okay, happy with that. Bring it my way. All tell you, Tom, let's say you the grower and say hey, Tom, I'm sorry. But this is not the quality we agreed on. And you should you should keep it for your own and avoid rejection. And by the way, that's also a good answer for you to grow because if you haven't hadn't received the sponsor, what would have happened is that you would ship it to the buyer and the ship, the buyer would reject it. And because it's already at the buyer, you would probably meet about 50 to 75% markdown, because it will take you a lot of time and effort to reroute it to another buyer. And that's usually demand on these kind of cases, we've clarified, you would avoid a rejection, and you can route it to a buyer that would be happy with the quality of the supply. So that's one way we create a transparency. But by the way, what's even more interesting is as long as this supply, which comes in the form of a pallet, or a box, or container, or whatever it is, as long as it has a barcode, right, its own unique barcode. And you keep this barcode along the supply chain down to the front store, the retail, actually, one can track the quality evolution of the supply from harvest down to the front store for a retailer and track what has happened to eat, and what was the shelf life. And there is a lot of valuable data that the players especially downstream, the retailers, and the old sailor can derive from that and optimize their operations.Tom Raftery:
Excellent, excellent. allowed, we're coming towards the end of the podcast Now, is there anything that I have not asked that you wish I had, or any topics we've not brought up that you think it's important for people to be aware of,Elad Mardix:
I think what's going to be interesting to play out is how quickly the industry is going to be receptive to this kind of technology. And the reason I mentioned it is because, remember, at the outset, I said, What was amazing to me when I was introduced to this industry, is I've never seen such an inefficient supply chain. Nevertheless, they haven't done a lot about quality evaluation, which is key in order to improve it. So what change why things are going to change now. And I think the reason actually, we're gonna see, and we're already seeing an accelerated shift into digitization, and automation of quality control in the fresh produce of supply chain is because of the following to the drivers. Driver. Number one is, I think, both Unfortunately, the COVID pandemic, which it's still with us, then we probably going to need to call him with it for at least a few more years. But also just the fact that people understand now, the risk of unsustainable food supply, right, with the current waste and inefficiencies we have. I think I think people in the supply chain understand that the only way to overcome the challenges that they have today, exacerbated by the climate pandemic and the food sustainability issues is by digitizing the supply chain and managing it in a much more efficient and objective way. In our technology. For example, one path to enable that. I think the other interesting trend that's taking place, which was not available even four years ago, is that four years ago, you just couldn't take a smartphone, any smartphone, Android or iOS device, take a picture and analyze the quality of the fruit or vegetables very quickly, you just didn't have those capabilities and availability. And today it is available, the smartphone cameras are strong enough. And the computer vision and machine learning capabilities are getting there, we are still working on building this database to teach the machine to analyze automatically. But we're already more or half the way towards completely. And I think those two drivers are going to allow a much more efficient technology adoption on the supply chain. And you know, at the end of the day, everything needs to come to translate into dollars, right. And I can tell you that our customers see today a very powerful ROI. The average payback period for my customers is less than six months for the cloud for the platform. And I think that's obviously a huge advantage. And also the fact that you know, we're a SaaS platform. So there is almost no upfront cost that a customer need to put at the outset of the purchase that would allow us to further penetrate the supply chain and hopefully, you know, help this industry reduce this 45% waste to something which is going to be much more manageable. valuable for the sustainability of the food supply chain global,Tom Raftery:
super Li. If people want to know more about yourself, or about Clary fruit or any of the things we discussed today, where would you have me direct them?Elad Mardix:
Sure. So first, everyone can go to our website www.cloudflare.com. And learn a little about what we're doing. But even more interesting, actually, any person in the world can go to our web, download our app, and perform a 30 day free trial with our mobile app. And so everyone can have the opportunity to kind of have me walk my talk and experience really the magic of glide foot through the computer vision. That's one path and the other path. I'm happy for anyone interested to learn more to email me directly at a lot of CLI for.com and a lot of these email@example.com and we'd love to have a conversation with potential customers. If the topic of your podcast imply we are fully integrated with most other software in the world, including SAP HANA earpiece system, then different w Ms. So it's almost like a plug and play to adopt clubfoot within the fresh produce QC arena for any customer today.Tom Raftery:
tremendous, tremendous. Hello, that's been fantastic. Thanks a million for coming on the podcast today. Thank you so much for having me and looking forward to staying in touch. Okay, we've come to the end of the show. Thanks, everyone for listening. If you'd like to know more about digital supply chains, head on over to sa p.com slash digital supply chain or, or simply drop me an email to Tom Raftery at sa p.com. If you'd like to show, please don't forget to subscribe to it and your podcast application of choice to get new episodes as soon as they're published. Also, please don't forget to rate and review the podcast. It really does help new people to find the show. Thanks. Catch you all next time.