 Good morning. Good afternoon. Good evening to everybody. Today, I'm happy to welcome you to the food risk predictions webinar. My name is Neil Marshall, and I have the pleasure of being the moderator for today's session. We have an interesting session for you today, very quite short and concise, but hopefully we can share some good interesting facts with you, particularly for people on the line and people who watch the recording. For people who want to know more about risk predictions and how to manage that without having 2530 data science engineers within your own team. Next slide please. I'm Neil Marshall, Managing Director of my own private consulting company now, previously, as people may know me from my roles at Coca Cola for 20 plus years. I have the pleasure today of welcoming Kelsey Robson from ABP food group, and my highlights Papa Constantinou from agronome. Apologies for I just butchered your pronunciation, but my Greek is not as good as my English. For Kelsey. I've just been talking to again recently but she has a very impressive background. Research studies have taken her from PhD to Masters to BSC from Seattle to the prestigious colleges of University College of Dublin. And also that our recent PhD at Queens University Belfast where she majored on agri food and land use so very pleased to have Kelsey with us to share our insights from ABP. And my highlights is really becoming a close friend of mine now, as he supports us so often on the data and information and educating us and leading his team agronome on how he discovers all these insights from the data and all the processes that's been done by that team. My highlight is also very well educated with a master's degree and a BSC from a prestigious university Athens, and he is really what I would call the brains behind the platform, along with Yanis. And he leads a team of engineers who help interrogate that data and create the right information for the customers. Okay, so the goal from today, what is it well we really want to try and as I said in the last slide take you through the risk predictions and how you can do that and what tools you need to to assess that data, you know there's so much information out there. But you need to be able to interpret that data and how do you interpret it, or you need a tool, why we're going to show you today but it's the best way without having a load of people on your staff with a lot of cost. But you need to be able to interpret and look for the right data so that's also critical. Next slide please. So where are we I mean one of the big challenges we're all facing now is the whole position of consumer trust. Consumers, if you see on the left side of the slide there, you can see that not everybody is trusted in the same way. And people generally would go for in the order of preference there first to the internet, then trust in probably the big companies and food business operators. But the most trusted people from a consumer perspective and obviously this can still vary by different markets is the government and the regulators. So one thing that we've been doing closely with that right now is working with these regulators as well, and with academia and with industry to also get the right insights to use the data. Because if you look them to the right hand side of the slide. Unfortunately, we still get too many food safety incidents from recalls and where people become ill or sick from eating incorrect food and 17% of the world's population is still way too much and everybody's intention is obviously try to drive that down much more. Next slide please. So why are we really talking today about risk prediction and risk prediction to me is the new way forward. We need to be much more inclusive of information and use it much more to drive decision making for the future. Industry tells us, you know, obviously, from my own personal experience, we're always trying to avoid a recall, we don't want to pull product backs very costly. We're not consuming and obviously gives a bad image to the companies who are doing it and the industry. If you have a recall and dairy company is going to affect the whole sector it's not just one company. So we need to be more proactive where we can and obviously everyone's trying to be proactive, but you end up being reactive because you're always chasing your tail if you're not careful. Another phrase that I use quite a lot is in that prevention area is trying to see around corners. So to see the future, and how can we do that well, we need tools, we need to use technology. On the right side again you can see, you know, there's so much data out there, and my high list can help us interrogate that further later. We have so much information it's hard to know which bit to use, and how can you see, you know, another English expression the wood from the trees how you know what's important, and how can you use that data to your advantage. But the biggest point as well is, when you're not investigating and following up and managing recalls, which is spending a lot of money on, you have the opportunity to be more proactive and save money and invest in different ways. The only way to do that in a much more beneficial way is to invest that money in technology to use and interrogate that data. Next slide please. So again, everything needs to be around science. I talked before about technology, technology can't work properly without people. So you need people to interrogate the data, and you need to be based on a scientific evidence that drives you to the right conclusions. Next slide please. And as we said, you know the data gets so complicated now and in the platform that we'll talk about and with Kelsey, there's so much information, it's difficult to know which piece to use and which piece not to use. You're doing your risk assessment you're doing hazard analysis you're looking at your suppliers, you're looking everywhere but how can you avoid those recalls. And some of the things that are going to steal some of this information from recent conversation I had with Chris Elliott at Queens University as well, you know, one of the big things that Chris was mentioning is the next risk for the food industry and the world basically is around climate change and the sustainability and the challenges that are coming. We've all just lived through the pandemic of COVID for the last 12 months. How many food issues have occurred, or probably less, there's been less recalls, but by numbers in the last 12 months, but why is that. Are we just storing up problems for the future. No one really knows yet, but it's probably likely that will be a spike in events coming forward. So we need to learn from the past to look forward so always looking back in the rear view mirror is not the right approach. We need to be more forward thinking. Next slide please. Yeah, and the whole point around information is what you do with that information. And how do you get the right information. How do you know you've got the right information. What are the countries or suppliers that pose you most risks and every company has a different situation. Everybody has different lenses which they use to interpret that data and experience obviously helps in some of these cases. But the biggest thing we need to move towards and I would suggest we need to strongly move towards is the use of data and technology to inform those decisions. Next slide please. So one of those things again, I mentioned Chris and I was talking to Chris again actually yesterday is the digital crystal ball so we've coined that phrase between us. This is how we use technology to look forward into the future. Next slide please. A critical point for us to use that data to use technology and that digital crystal ball to interpret the information. You need to be able to have a simple dashboard where you can assess those risks. Look at the different countries, capture into one place. What better than to go to one place every morning check on you as you log on to your laptop looking one place to see all that information. I know in my case in my previous role. That's what I like to do. I like to see everything review it, move on, allocate the work out to the people that need to look at it. But you need to use that data and that interpretation to be constantly checking and wherever possible, we need to automate those approaches so it's not relying on people and experience. It needs to be built into a exacting process that looks through the data using predictive analytics and the technology behind it and the AI with the algorithms to make sure we are using predictive science to look for the risks ahead. So in this webinar, we're going to talk about that more and get some practical experience from from Kelsey from experience. And also then my list is going to take us through some details and analysis, and then also a demonstration of the food archive solution. I think we have a poll. Yes, just before I pass to Kelsey, we just like to ask you to answer the poll, please online, click one of the responses, and then we'll incorporate that into the answers later on and share their feedback. Okay, next slide. So I think at that point I'm now going to hand over to Kelsey. Okay, so before we get quickly we've got the responses so we have the responses so maybe this is a question into you, Kelsey. So let's just have a look what do we get we get probably 40 to 46% saying merging risks for ingredients is the thing that most people want to see along with the suppliers as a joint time. So I don't know about you Kelsey so what would you like a crystal ball for to see the risks ABP. Thank you for the intro for you for your first sense and good morning, good afternoon. Yes, that's a good segue. Hello everyone I'm Kelsey Robson again I work with ABP beef manufacturer. I primarily located in Ireland in UK, and we have been involved with agronome for probably the last six months, working on a project and predictive analytics because we do see a lot of people in knowing what's next, what are the emerging risks, what suppliers what areas are issues going to come up in and the big question is just where do we need to focus our resources. We act proactively particularly when it comes to food fraud where there's so many unknowns. Where can we manipulate our mitigation and prevention techniques and testing so that we are the most prepared for any upcoming issues, whatever they may be and if we have a crystal ball to tell us, focus all your testing or focus more of your testing in this area, that would be so helpful in how we can approach emerging threats and emerging hazards. So, next slide, please. So, then going into what we want to achieve with this project and there are a lot of different things a lot of different factors we hope to achieve I think with what we why we got involved are the big picture issues we want a crystal ball we want to know the emerging risks, but then there are a lot of little questions to get there again what factors contribute to food fraud do we need to look at prices and trade and climate change and legislation, how do those factors contribute to risk and also how do those factors interact with each other, and based off of those interactions, can we get that crystal ball to see what types of food fraud are coming and what are the emerging hazards in our supply chain if there's a climate issue, is that going to cause more substitution issues or is that going to cause more issues in mislabeling things traceability so having all that information in one location where we could quickly look at it don't need teams of people to sort through horizon scanning and media mining but have it in one area where we can quickly see patterns occurring and know what those patterns mean and some algorithms that tells us this is the hazard coming it's coming in six months time or it's coming in a year's time so that we can adjust our testing methods to to protect our supply base. So overall, we just hope that working with agronome can give us this crystal ball and help us wear it to help us mitigate these issues. Next slide please. So that brings me to our last point what value do we see and the value is knowing where the issues are coming so that we can adjust our mitigation techniques and keep our supply chain and our supply base safe from any upcoming hazards and any upcoming threats to our supply base from around the world. So, thank you. I'm sorry I'm having a bit of a delay with the technology here so I didn't mention at the beginning I'm here in the US. Kelsey's obviously in the island, and then the agronome team are in Athens so we're bridging all different time zones and continents here so that's that's a good thing. And that's the reason also we're doing next slide please sorry just so people know this, we don't want too much of a delay. And again Kelsey interested in size so maybe even how we can say wait in the hairless. And he can tell us now more. How can technology help us and how did it help you with ABP. So my house. Perfect. I thank you very much Neil for the introduction and by the way congratulations on the attempt to pronounce my last name I know it's the Greek one so thank you. I know in order to talk about how technology can help us and perhaps you can start with a bit of an introduction to our new show if we move on to the next slide. Who are we who is agronome. We are the food safety intelligence component. What we want to do is that we want to predict food risk in order to make prevention a data informed decision. And if we move on to the next slide. Now, more specifically, the reason and the reason behind what we want to do and what we want to accomplish is that we want to empower food safety professionals out there to help them with data insights in order to help us gain access to safer and healthier food. So we move on. The next slide now, the main reason we're talking about data load here so it's actually one of the reasons that companies tend to agronome. One of the reasons that actually we got in contact with ABP and CalCIS team. We're in contact with us because we're a data company we know about data we've been doing data for quite some time now. And they tend to us because, okay, as you most probably the ones in the audience know this better than me that the food safety sector is a sector that deals with various data types, many, many different data types we have food records. We have border rejections, we have trade data production data, there are millions and millions of data points out there, and all of them are in a multilingual way in various languages, various formats, they are highly heterogeneous data points. So, what we know how to do and what we've been doing over the past years that we attempt to make sense of this data collect, transform and translate this data, and then to make sense of them and help the companies that tend to us in this way. And the next slide. Now, how do we help them, we have two offerings here on the one side on the left side you can see food archive our software service solution, and this solution focused on risk assessment and prediction. And by doing about risk assessment we're collecting data, going back many years, we perform risk assessment on ingredients on hazards on countries and continents, but we move it a step further, and you will see this in a little while during the interactive demo. We move also to a predicted, to a predicted state so we know what has happened in the past, can we use this data in order to attempt to predict what will take place over the next 12 months. This is a software service solution, but as we mentioned previously, we're also heavily dependent on data, we've been doing data for quite some time so we should have a data service solution then this is what you see on the right of your screen, risk data. And now this is our data service solution this was actually announced today just a couple of hours ago by our CTO in a previous session. It's data as a service concerning food safety and food risk. And this is what we offer but if we move on to the next slide. Now this is just a quick overview as far as the data points that we collect and process and keep in mind that this data records that you see as you can see, there are millions of them. These are the records that are constantly evolving. So over every five minutes, we are collecting data from resources that we keep track of and you can see here and over just follow the number of sources that we keep track of. You see the global coverage of this data almost the whole world discovered. And the last point here is the data types know this is something really interesting and again I'm sure that many in the audience know this way better than me that the food safety sector has a lot of different data types this is the variety as far as big data is concerned but there are not many data types and if we move on to the next slide. We get just a quick overview as far as the data types involved and collected in our systems. So, apart from food records and border rejections, you can see that there are also lab tests, there are social readings coming from my farms who have production data trade data social media data food safety news announced on websites all around the world. And here you see mentioned before legislation places and the respective updates and my limits inspections and common price and so on. You can see here just a quick overview as far as the overall numbers are concerned. If we move on to the next slide. Now, let's turn our attention to the tailor made prediction case study that we did with Kelsey and her team. And if we go on, let's start with what we wanted to answer there. The question was that, okay, we mentioned the variety of data types in the food safety sector. Which of them playing important factor in order to be able to predict what will take place in the future. We have all these data points. They are, they are of various types. And we identify which of them will take place. So, we will help us understand what will take place in the future. So we have a business question, then we have to choose the data. And in this case, although we experimented a lot and Kelsey's and her team suggestions actually proved valuable to this. We identified the following data types that were of help in order to predict what will happen. We involved incidents in this case, lab results, production data, trade data, and price date. And bear in mind that this specific tailor made use case that we did with Kelsey and her team was in order to be able to predict food fraud cases happening in beef in the future. So we have the data, we have the question we have what we want to accomplish. And of course the missing piece would be to identify the prediction method we will use in order to do this. And this is just to give you a quick overview. And as far as the use cases concerned. So, what we did is that we deployed we trained and deployed a dedicated model a dedicated machine learning model for the case of MVP in this tailor made the tool that we developed. Now of course, since this is actually tailor made to a big case where unable to show a live demo but we have here a quick overview on the image on the right. And how they, they, they are so in the start here in the bar scholars, and the specific case that we use doesn't evaluation step was actually a fraud outbreak involving Brazilian beef. Back in 2017, there are various things we can talk about but please bear in mind that the approach that we used in this case, actually was a correlation one so mainly you can understand this as input in many different data types, many different features of the machine learning identifying the correlation how does one affect the other in order to be able to a assess what will, how many incidents will take place over the next month. And how will this affect the risk assessment for a, in this case, for all happening in brief. Again, I mentioned here that this is a specific tailor made use case. So that really like now as you can see a quick overview as far as to the actual data records involved in this specific use case so you can see here that apart from for cases and we've the other data types are actually in the scope of millions of records or hundreds of thousands of records. So let's go on to the next slide now. This is what we did. As far as the case, the tailor made case for ABP. However, I will mention this is a tailor made case so we worked heavily over the past six months, which with Kelsey and her team, identifying new, new data records, new data sources, and running and running training and retraining our models in order to be able to have a production available. This is one thing. And on the other side, we have food a guy. Now, this is accessible to all this is, as we mentioned our software as a solution. And it's a food risk prediction platform. And what we do, if we move on to the next slide, by using food a guy, we, we have a promise to our end users that what we do is that we'll provide this predictions for all and this is available to every user of food a guy the prediction part that will be given to interact the interactive demo a little while. And in this case we're covering all the important ingredients throughout everyone's supply chain. If we move on to the next slide. Now, how do we do this how does this risk prediction take place. It's split into various steps. Of course, on the first step we have the risk identification but we have to identify what is the risk that will take place and this is either on either an ingredient level or supplier level, because of course, in this case, countless of fortune play different factor. And, of course, the end game is to actually be able to inform our users and users before something takes place in their supply chain and alert them previously so they can perform some medial steps. Next slide. Now, as far as the first step of the risk identification. Now, this is where the data comes in, we collect data and as we mentioned before, we collect them up every five minutes. And this data are global food safety data food recalls and border rejections happen in the market level. And by having access to this data, we're able to identify in magic hazard so which ingredient is so is so in an increasing tendency in terms of number of incidents or for specific hazards. And in this data, we also have available suppliers name of the food companies that have either manufactured the recording gradient that may have imported the recording gradient pocket state. So we have information of food commands. So the way that we use this data in order to evaluate the surprise that we take this into account, and we're able to perform supplier assessment on top so if we move on to the next part. Now, how, how does this help, and how does all of this all of this collection of data here in order to identify it. So we're going to get out here for specific cases. We have February next back in 2017, the ethylene oxide outbreak, Prochloris mandarin in manders and oranges, and the latest in peppercorn so they were all caught by a scan part of the solution. If we move on to the next step. Now, if we move on to the next slide, please. So now having access to all this data, we're able to perform risk assessment and using this data we can identify where what is the risk value of specific ingredient. And if we move on to the next step. Then we can perform the actual prediction on top. So having all this data performing the analysis, we are able to to attempt to predict what will take place in the future. And this is what is covered by food guy. Now, during the interactive demo, we will show you the prediction dashboard but just a quick overview as far as the other functionalities of food guy. They have to do with the scanning part so we are constantly monitoring for new cases having worldwide and we are later and users for them. So if I imagine hazards and perform the respective analysis performs applied evaluation. We allow our users to perform reporting in an automated way. We collect and allow these insights available in the global lab testing sites are available in our system. And as mentioned in the previous slides, the risk assessment and the prediction part are available there. I think it's time for them. Thank you, Myles. Yeah, it's good. So I think one of the things before we go to a demo is, you know, it's just to summarize the many there are many assets and points available within the platform. You can see the seven items there. So, you know, the quick poll question now again please if you could just submit your answer which of the food guy modules would you more interested in and be more valuable to you. And after we just get the quick response I'm going to give Myles a little bit of a challenge and he's going to give us a bit of a 15 minute demo so it won't be too hard. I think for Yanis I like to give him really hard questions so I'm more friendly with you, Myles, so I'll give you easier. I hope Yanis is listening as well by the way. So what do we get? Do we get the answer if any? A few more minutes. Second, sorry. There we go. So what do we have Yanis? We have risk prediction. What a surprise. That's always mine as well but usually this is also a close tie between supplier evaluation as well so maybe if you can look at that and I know from the polling and descriptions earlier from the feedback people are looking for information around peanuts Myles so maybe if you can do a deep dive in the demo into peanuts and then how can that help us with the risk ranking I think that would be useful. Thank you for the demo. I might jump in at some point and ask you some questions so just to keep you on your toes. Feel free to do so. Okay. Now let me quickly share my screen. Hopefully everyone can see my screen. Perfect. Yes. So before we go into the actual prediction and more specifically we will use opinions for this. Let me quickly mention at this point that the prediction dashboard within for the guy, along with mostly in almost all of our other modules available and actually customizable so everything you see in this prediction dashboard is based on choices. I've made and each user has made under the customization step. So in this case, I've created a bread preference prediction preference that involves as a counter India and the specific ingredients cereal, peanuts ice cream and ginger. But enough with the overview. Let's dive into the actual predictions. Now, I'm switching over modules and clicking on the prediction dashboard. As you can see, the selected ingredients under the customization step are available here. Now, as you very much correctly mentioned, we had a poll previously on peanuts was the ingredients, the ingredients that actually draw the attention from most of the attendance here within within our system and in this webinar. I'm selecting peanuts in this case and let's dive into what is available in our in our prediction dashboard. Now, first up, we have a quick overview as far as the total number of incidents go for peanuts. So, mainly, what we do here and the overall approach. Let me just quickly mention this again. We have a lot of data going back for the years. And the main idea is that can we can we use this data. Can we take advantage of the seasonality behind all these time series involving ingredients involved different different hazards, or different models and we'll take advantage of all these trends and seasonality and be able to train models train dedicated model for each ingredient and hazard in order to identify what will take place over the next 12 months. And what do you see here are actually the results of our most accurate and predictive capabilities so predictive models. So, going back to the actual dashboard, a quick overview as far as the total number of incidents, what took place in the last 12 months, what are most accurate models believe will take place over the next 12 months, and what is the expected And if we move on to the chart on the right. Okay, so we know 133 cases will take place for peanuts based on our most accurate models. But how but how are these cases spread throughout the next 12 months and you can see this on chat here if you hover on any of the points on the red dot on the yellow dot line, you can see the respective values. Okay, we've got a quick overview, can we dive in deeper, which are the specific hazards that are likely to increase based on the outcomes of our most accurate models. And this is what we attempt to identify in the next table over here. We're still looking into peanuts, and we dive in deeper so what we take place for peanuts, what are the hazards that are likely to increase over the next 12 months. In this case, we have mycotoxin and aflatoxin, no surprise there, but also the cases of pesticides or options of health certificates are also available. And again, what signified here is what took place the last 12 months and what are most accurate models believe will take place over the next. For instance, for the first line specifically for mycotoxin appearing in peanuts. And if we have this in OLEDs, if we have this estimation, as far as to what we believe will take place over the next 12 months, then we can also perform risk assessment in the future. So, this is what we signify with this blog on the right, the current snapshot under the actual tab we have the current snapshot of the risk. So you can see it here in a color way based on for the guys risk assessment formula, what is the actual the current snapshot of risk for peanuts, and how will this value evolve over the next 12 months based on the prediction we've made. And you can see some numbers and expect the values changing here. And now that we tapped into risk, how will this evolve because keep in mind that what we do is on a monthly basis. So we're talking about 12 months in the future, as far as incidents are concerned, and on a hazard level, and the same thing we're also doing for the risk so you can see here on the next chart, the risk evolution for the risk assessment of peanuts over the next 12 months on a monthly basis. And they will cover you as far as to when the sharpest increase will take place over the next months. Now, as we mentioned, the food that guy is a full customizable system so in case you bought the product recipe that involves this ingredient, you will see it appearing here along with the estimated risk in the future. So the product, other ingredients, and throughout the expected vocabulary that we'll use that will be affected by them. The prevalent hazard. And finally, this is where the counter of interest that was chosen under the estimation step comes into play as far as predictions are concerned. If you have a distribution, a predict a distribution as far as the predicted number of cases for the counter continent of interest, in this case, India, what are most accurate models believe will take place over the next 12 months, as opposed to what has taken place over the past 12 months. And this actually concludes the part of the prediction, the prediction dashboard. I believe me you mentioned that the second result had to do with the supplier evaluation. Yes, yeah, yes also so yeah maybe you can dig into the supply piece and show us shows that as well as I think that's really interesting and you can click down and follow the trends and use that. But I think the other point to explain is the supplier assessment piece you can also use in the dashboard. However, and we were actually prepared for this just in case this module is called highest, as you can understand for anonymity purposes, although what you see our screens are taken from a system directly. Unfortunately for an independent purposes were not able to with not want to show you a specific food company. So we will do this in an anonymous way over the next slides. Now, this is a natural. Another point on my list as well you know that what especially when you talked about the data feed as well the API. No individual companies data is secure. It's all in the cloud, and it's segregated from other people, all your other information is anonymized in but if you were a user. If it was my company I wanted to go in and enter my supply names in here. You can do that. Yeah, I know you can't show it now because for confidentiality but you can enter information there, and you can pull out specific information about current suppliers, or future suppliers. Sorry, I'm doing your demo for you. And then give you the information. Yeah, I'm giving the information about you know if I want to source from that supplier in India. What are their previous issues cannot pull information from the system that I wouldn't be able to find from the normal Google. So maybe you can just talk to that as well a little bit. How do you get that how do you then make sure it's correct the data. What do you and your team do to make sure it's right. This is why we've actually now the slide you're currently seeing is a screenshot taken from our system. What we want to do is give our users a Google like look and free and be able for them to search throughout our database for any kind of supplier that we want. So in this case we're just searching for supplier a and then I'm sure you can understand the name of food company may appear in various formats. So in this case we're setting for supplier a but country a is referring to it as supplier a incorporated or LLC, or any other vice name so what we do in order for our system to respond with a respective results. What we do is that we aggregate this information together. So this information may be alternative names, how well the food company is called throughout the world, but also we collect information about subsidiaries so by setting for a specific supplier. You can actually go over and see information that also aggregates information coming from their subsidiaries. This is what we showcase over the next slide so this is how you can search for supplier. Just by clicking on the supply check you access all of the data that we have collected going back for the years in total. Under the my suppliers tab we can perform specific assessment for specific suppliers. If we move on to the next one. We can create a library profile based on the data we have collected for supplier in this case, and you're seeing a respective screenshot from the supply profile of supplier a. So let's move on what else is available here is that let's say you've been putting money supplies and you want to get a quick overview as far as a risk is concerned for these suppliers. We have a dedicated dashboard for this, this is where all the suppliers you've added under my suppliers tab under the customization phase, you will see them appearing here as supplier a BC and so on. These factors coming into play so we have incidence risk, ingredients risk, counter risk, recalls for the rejections inspection, scrolling letters, and so on. And you can also add your own data as I mentioned, you can add your own data that will actually be dedicated to your instance of for the guy. And if we move on by adding your data you can change also the weights and how we calculate the risk for its supply. And this is all cases in this slide. We're going and going about it you in this case. Thank you my house yeah so obviously we can show everything as clearly as we'd like to. I think one thing to just be aware of, you know you can obviously follow up separately and I'll give you a name and a contact in a minute. If you want to memorize and just go through what we've talked about. Basically there's three different options that you can choose from to select with the food archive package from starting to premium to diamond and different features are included. Obviously it's all there but the pricing depends upon which elements you need and which are most suitable to you. Next slide please and the point really from that is, you know, if you want to get a risk prediction plan customized to you, please follow the link, capture that link, or use the QR code and contact at the team and the customer team. But I think just just to summarize because I think we're running out of time as always seems the case with these webinars. Now we've had an interesting overview from the demo from the tool from Alice, we heard from Kelsey, and thank you again for your insights and how you started to work with agronaut from ABP, and particularly around the B for us and the food fraud issues to mitigate those for your company. And then I shared some insights from my side about how I've used the platform, both in previous roles and also now working with agronaut to further develop and enhance the tool. So if you want to use your digital crystal ball to get more data to automate your processes to take away some of that manual risk assessment that you haven't to do by people, and you want to prevent some of that investment chasing down recalls and reacting We need to use the data more to enhance the processes and use that data science based data to make predictions for the future. And this is a great platform and a great tool to help you do that. I think my final just comments really about agronaut as a company and Nicos, the CEO and his team. They are really agile, agile in a good way that they adjust their very keen to help and learn and work with you. And as I'm highly mentioned before, you know, Janice announced at the GF side conference this morning about API risk of data approach the new, new solution they're always coming up with innovations to help and enhance for innovations that can help you do your, do your work better and enable your And the whole use of digitized technology is critical for the future for the food industry so I would say, you know, contact agronaut, use this number and use data for yourselves but thank you again. We'll answer the questions that came up in the chat I think Amazon has already answered some of those already, but we'll follow up with any questions specifically because I know we're out of time there. Thank you, my list. Thank you, Kelsey, and thank you to the team. Thank you. Thank you everyone for attending. Thanks for attending. Yeah, thank you.