WEBINAR: Developing Insights with AI

Alright. Hello, everyone, and welcome to today’s webinar, developing insights with AI. I see there’s people trickling in, So allow me to first introduce the session before I hand it off to today’s speakers. On behalf of the MarketLogic team, my name is Carolyn. I’d like to welcome you all to today’s event. However, we are not just speaking as Market Logic. I’m happy to be joined today by our friends and partners in the insights industry. Joining us today, we have Mario Coletti of Nextatlas. We have Sjoerd Koornstra of House of Insights, and we have Guillaume Aimetti of Inspirient. Before I move on to the agenda for today’s show, a bit of admin up top. You’ll see that there’s a questions tab in the bottom right hand of your screen. Feel free to use that throughout the show. If, we’re gonna be saving the Q and A section to the end. So as speakers are talking, if there’s a question that you want one of them specifically to answer, just include their name in the question. I’ll be sure to direct that question to them when the time comes. If we don’t get to your question, we will, of course, answer it offline. And if you need to dip out early, we will be sharing this as a recording tomorrow so you can watch it and rewatch it, at your own time. Insights. For today’s show, we’ll be diving into again developing insights. I know there’s been a lot of discussion in our industry about how AI is making insights more accessible, how we’re allowing people to act on insights. But today we wanna do a deep dive into how AI is helping us develop those insights in the first place. Specifically, we’ll be looking at trend forecasting, data marketplace analysis, and of course, survey analytics. These are all crucial parts of our jobs and often take a lot of time. So I’m happy to be joined by an expert field of panelists and presenters to help us navigate this. Thank you once again to Mario, Olaf, Sjoerd, and Guillaume. And with without further ado, I will hand it over to Olaf, our first speaker of today, to start things off. Yes. Thank you, Caroline, and, welcome from my end. My name is Olaf Lenzmann. I’m responsible for innovation and product here at Market Logic. And I wanna open today’s session by talking to you about how we can, answer business questions with AI with our Insights AI as a as a foundational capability. And then we’ll see with our other speakers how that also can extend, into other areas to further enhance and advance those capabilities. But to begin with, answering business questions, well, the fundamental and the underlying business problem that we’re tackling is very simple, and it is this. You have an important business question, and now, well, you’re faced with either, if you really wanna do your job right, spending a lot of time, maybe hours searching for the information, the knowledge, the insights that you have, turning all the stones, looking at all the reports, scanning all the material, finding all the information to make sure you really have the best and most relevant, recent, up to date and accurate understanding, of what you know as a company about the market customers and consumers to answer that question. Or maybe you’re on a lot of pressure. There’s not the time to do this. There are other projects and parties, so you have to make some kind of an educated guess, meaning maybe looking at a couple of pieces of material, but never having the time and the luxury, so to speak, to look at everything and fully, leverage the entire knowledge that you have to answer that question. And this is what we do with DeepSights. DeepSights is our AI product that reads all your research, all your information, public, page, in house, syndicated, and that ingest that inspected and that gives you the answer to the business question. So then that you can be assured you have the real and the best answer that you and your company has in response to your question, and you have that within seconds or a couple minutes at most and not within hours or even days. And let me just speak, have the product speak for itself. Let me bring it up and and show you in a very quick three minute demo what it does and how it works. And now, looking at, looking at the product, if we go over here, what you see is we have a very, very simple interface where you or a stakeholder can go. You You can put in your question in very natural language. What is your business question? And now the system goes in, draws all the knowledge, all the insights, all the reports, all the research, all the internal and the external material that is available, evaluates that, inspects it, weights it, and then picks out the most relevant and most recent pieces of information and gives you generates you the answer to your question based grounded in that information. And it does that always referencing directly back to the underlying source to the exact spot where the information was pulled from. So to give you the reassurance and the transparency and the ability to verify, but also to learn more and look deeper. And it gives you not only one question, one answer to your question. It can give you multiple of those to also reflect the fact that there’ll be multiple lenses of answers to your question. And you have multiple sources that may be also, of course, recurring waves of reports, for example, where you might wanna see more longitudinal, answers. So you can easily scan those, answers. You can consume them. You can share them. And you can get a very quick first impression of what is it that we know about this question and topic. But then also you can ask the system to go in and do another pass, look deeper, look further, take a minute or two to do a more elaborate analysis and come back with a full report of everything from all the sources that are relevant. And while this may take now a minute or two, I, of course, have prepared that. And so I can show you what it does such a report look like. What the system will come back after one or two minutes to you is then a fully synthesized AI driven report, compiling all the relevant sources, still retaining the full transparency of all the quotations of all the material, bringing it into a structure that is reflecting the content of what it could find and not only referring to internal material like we see here in some of the reports that were cited and quoted, but also, for example, ingesting public content, looking in news feeds, looking at other sources, so giving multiple angles while always, again, clearly flagging where content comes from, making that very transparent. And then as a user, you can simply download also this report very quickly and easily for offline takeaway and for further refinement for downstream work to then produce the final end product. This is in a nutshell what DeepSights does. And going back to the presentation, we see now with Insights, we’re able to provide those trusted answers to the business questions that are grounded in your your your knowledge and insights and your analysis information as well as public and then syndicated content. And DeepSights is available around the clock. You can talk to it through the user interface we saw in the web browser. You can talk to it through Microsoft Teams and chat with it and mention it with your questions or with Google chat or Slack. You can send it an email with your question, or you can also use it and inquire it through APIs. So to integrate it into other systems that they themselves may require insights about consumers, markets, etcetera. Now a common question is, that’s all great, but can I just use Microsoft Copilot or any other AI assistant that’s out there? And while these are, of course, wonderful pieces of technology and very capable, they are not really tuned to this use case. Just to illustrate that, let’s look at the typical specimen of a piece of information that is the foundation, the grounding of the answers that we give. And I just picked something from a public study, but that is a very representative way how you would find information in such studies, of course. And then you have to ask yourself as the AI is attempting to give you an answer based on, for example, this page, which might be relevant to your question. The question is, does the AI understand, well, what market was this, researched in? Is is this really applicable or transferable to my market? Does it is it able to make that connection? Is it able to, understand when this was researched? How it was researched? Is this a reliable source? Is this a robust piece of information? Does it really weigh whether this is the most relevant and recent data we have on the topic? How does it weigh at all? What is relevant versus what is reason to the specific questions context that I have? Does it consider this, in terms of my best practices that I may have in my organization? Maybe we have authoritative sources we picked and defined for certain kind of information. Is it able to understand that and reflect that and adhere to it? So that, for example, if I have questions about, let’s say, certain, market share numbers, then it always picks that one source that we wanna rely on, or will it just pick whatever talks about market share? Can it even read the charts that we have here, or can it only look at the text? And if it only looks at the text, is it able to make sense for just the text, or will it maybe give me wrong answers because it will confuse the numbers and the labels that we see here? And does it understand what really is the difference between all those elements it reads here, whether something is really a research finding or just a quote of what somebody said or where that was a questionnaire question. And it’s again important, of course, to understand the context of this piece of text to be able to decide whether you, as the AI, can give an answer that is relevant and really reflecting on the actual research or that maybe mistakes something that was a research question and a questionnaire for a ground truth. And ultimately, is it able to say I don’t know if it doesn’t have an answer? That’s a common problem with with those, AI tools. They will tend to always give you an answer, and if they don’t have the evidence, they will tend to make it up as well. And this is where we really have invested a lot of effort, a lot of work with customers in lead client relationships. We have been on the market for more than a year now with the product. We checked most of these, and we have also now in the upcoming updates this quarter, further extensions to to tweak, tailor, the behavior to your requirements and also to fully understand visual information in all the material. But then, of course, it’s not only about the assistant. It’s not only about the AI. It’s also about all the information that is needed to feed it. And there we have, like we’ll see in a moment, with our partners, out of the box integrations. We have ways to get information in from all, angles in a three sixty view, which is needed to unlock the real value. So if you’re interested, of course, contact us for a free trial to see how it works to convince yourself that it does the right thing and that it can properly appraise the nuances of market research. And with that, I wanna segue into the next segment where we want to talk about one very specific and extremely important source for information, for insights that needs to go into the organization that is needed to answer questions, and that is insights and trends. And here, I’m joined by Mario from Nextatlas who will take us now through what their solution is and how we also then can bring it into this bigger picture picture of answering questions. Alright. Great. Thanks. Thanks, Alex. And hi, everyone. Nice to be here with you and, have the opportunity to, present together with Market Logic and Insights, what, we are planning to do and, bring it to the DeepSights knowledge base. So the sorry. I got a warning failed to switch all your device. Please refresh the page to take effect. I don’t know if, if you can hear me. Hopefully, you can. Yes. Okay. Sorry. The warning came up. So, so the, one of the key challenges that, many clients have is, you know, being able to predict in a relevant way, what will happen in the future for their business, how to define and bring insights to the core of the strategy development, but in a way that is relevant to their business. There are so many different sources of trends, from, different, opinion leaders and, also different analysts. The challenge with, all the trends, content that is available is that, usually, is very macro. It is also, often, not tracked and not, continuously monitored. And, therefore, there are provision or or a foresight about what will happen in the future, but, those are predictions that are happening, and then, nothing has has happened afterwards to this to be able to say that, eventually, those predictions are happening, really or not. Now everything, has changed with, with Nexatlas from a point of view of trend foresight because Nexatlas, is is a platform that we have developed in order to be able to, first and foremost, identify from the beginning what may be weak signals that may become, mainstream trends in the future. But the other point of, using artificial intelligence technology that we have developed is the ability not only to identify and say which will be the trends that would be relevant for a certain industry or category, but also the ability to track that, over time to see effectively, what is happening and how things are evolving. How we have done it. So, basically, we have developed the technology in our company, which is a mix of data scientists and insights strategies that, have developed, as I said, the capability to identify in the market, of and in the space of social media what we call innovators and early adopters. We have, identified a dynamic community of over three hundred thousand early adopters that are active in social media and specifically in Instagram, x, YouTube, and and Reddit. And we follow what they say and how eventually, they are talking about the their, passion categories. It could be food. It could be drinks. It could be fitness. It could be health. It could be technology, whatever. Through the analysis of and the continuous analysis of, millions of posts that are identified and found from those three hundred thousand early adopters and their followers, we, have identified and developed a technology to really spot what, we call the weak signals that I mentioned before. Those weak signals that may become mainstream over time. Now the challenge of, identifying a weak signal and predicting that will become a a trend, it is all based on tracking, that kind of signal and see how it grows and changes. So the tracking technology is the one that allow us to really identify what is a fud to where what is a trim a a trend and if a trend is a microtrend or a macrotrend and the the the industry that is related to. Now, obviously, this is where, the correlation at the where the the partnership with Insights come into the in coming into place because by crossing and combining the data, that are coming from, everything that is coming from Market Logic together with Nextatlas, The analysis can actually become, and bring relevance to, the specific category or the specific industry. The uniqueness of our technology is that, we, are able to analyze and create and understand correlation between visual and verbal content. And this is why we analyze a mix of different social media, platforms such as Instagram, x, Twitter, and sorry, x and and YouTube and Reddit simply because they are they are either reaching text and verbal or reaching visual content, both static and dynamic, and therefore, pictures and, and videos as well. By crossing and combining NLP techniques techniques with computer vision models, we are able to extract value from text images and generate insights, understand, in in-depth what consumers are talking about, the space and the kind of and and the profiles of the consumers, sentiment, and many other aspects that often, the the trend, directions and the and the trend the content. As I mentioned, the key, benefit of Nextatlas is not just only on the prediction of trends, but the fact that we can constantly track trends and therefore analyze the speed of evolution of trend either growing or declining. And therefore, we track the future trends and the behavioral shift, delivering both short, medium, and long term professionals, and we keep them constantly up to date, and as the the platform is always on, and, the predictions and the updates are done on a weekly basis. As I mentioned, it is important to combine multiple sources, both visual and verbal, and, also that, sources that are able to cover all kind of, generational, groups of, of consumers. So we are able both to analyze what are the trends that are relevant for, the youngest generation, generation y, and then x, and then the millennials and the senior people. And, as we now, we can say that also senior are highly active in social media. We, basically, can identify, all across all target groups, exactly what are the key trends that are relevant for a specific industry. In essence, through the site, it’s possible to access, the Nextatlas platform. And, obviously, the, kind of queries that, all of, presented and demoed, will and are enriched by the Nextatlas content once once the Nextatlas, platform is integrated into the DeepSights product. The objective of the platform is to detect the emerging trend and understand on micro and micro level the changes that have impact in in different markets. As I explained, it’s based on our community of, over three hundred thousand industry specific innovators, And the the the platform is a very rich repository of trend, fact, and insights that have been, collected, for more than five years, and therefore and therefore, we have been tracking and checking the evolution and the changes of those trends, over time. The the constant work that is done by, the the platform is to scout and analyze the data, that are coming, at industry level and, combine and analyze the correlation across the different, social media platform. The use of, AI machine learning content and visual analysis, and so on is a, is, is allowing us to generate automated insights, that then are further announced with the expert creation. So we have an expert team that is validating and checking constantly what is happening and what is being published by the platform. And then the platform gets constantly, updated, so that, eventually, what is published, is, and can be seen in a very easy, to use format. So in essence, we start from, what we call unstructured data. And unstructured data, are millions of data that, are identified as weak signals and, give us a very much a broad view about, key changes that are happening, in social media conversation related to topics. And here you see some example of topics, ingredients, homemade flavor, and so on. At that stage, in that format, those unstructured data do not mean much. But once they are organized and, and and grouped and, obviously, the correlation, is, is done across, different consumer groups and so on and so forth, they can, be grouped into, trends that are detected and, better organized organized so we can actually lead into, in into some specific insights. And then those specific insights are then supported, by, statistical data, such as trend lines, and many other information that are contained in the platform. The ultimate objective is that from unstructured data, we get to an output which is very easy to use and navigate and user friendly for for, for the users. And as you can see from this image here, I mean, we try to visualize and create dashboard, which are nice to see, but also very informative, and, provide the key information that, the user may want to know about specific, a specific trend. Now Caroline will play a little demo video of, of the platform, and I will, basically give you, a quick explanation about once you access the platform, what you get. So what you see here is the first page of Nextatlas, and the first page of Nextatlas is basically feeding, to the user, a number of different, information or suggestion about what are the latest trend or what are the latest factor that, may be interesting. As I mentioned, the platform is, very rich, contains over six hundred trends. And in this page, you can see that there are hundred and sixty trends, which are related to, a specific profile of a user. So you can actually tailor and, and and personalize your platform to your needs. So only to, get, the insights, the trends, and the facts that, that you need. Once you have done your personalizations, then, you get the, specific information which are relevant for your profile and what you are looking for. And then, obviously, the platform suggest you to get into insights and trend that, you want to explore more. Once you click into any of those icon, for trends or insights, then you get into the specific details of a trend or an insight. And this is a demo of, the trend that I was showing you earlier on. As you can see, navigating the trend, you see clearly what is the, evolution over time, what are the different geographies, the key different target groups that may be interested for the trend, or related to the trend. You get concept and tags, which are very important for your social media strategy. If you are, pushing or targeting those kind of consumer groups with some proposition that is related to the trend that you see here, which are the keywords that you needed to use in order to increase your SEO. You understand better the occasions, that are related to the trend as well as the industries or the potential brands, that maybe that maybe engage or, have a higher relevant, correlation. And, also, you get a highly visual kind of, examples of, the the trend that justify or, basically, give you the sense of, what the trend means, but also give you an indication about, if you need to communicate to, your consumers and, you need to relate your proposition to that trend, which kind of imagery, you need to use. You get, sentiment, emotions, sources, and other, key components. All these kind of, information that, are there can be downloaded in PowerPoint, in PDF, and so on. So, basically, the platform, beyond the fact that it’s very easy to use, and can provide very, very, rich information, in a in a very quick time, helping a lot on, let’s say, combining your, the trend as aspect with the with the strategy that you’re planning to do. I hope that they this gave you a very good, kind of, demonstration and, overview of what can, Nextatlas do. Obviously, integrated with with Insights and all the, data from DeepSights. Out of that, you can actually get a very, very rich content. So without further ado, I will hand over to Olaf that, also will explain how the Nextatlas gets integrated into, DeepSights. Yes. Thank you. Thank you, Mario. Thanks for the excellent overview of your platform, which I think is really exciting and, of course, even better. We think is that Nextatlas is is fully integrated with DeepSights. So if you recall, as I as I ran a report earlier in the demo about a question and as the system went out to look at all the knowledge and insights and data points accessible, it would would also now be able to automatically inquire with the NextEI plus platform on consumer trends that are available on that platform where NextEI plus platform in response generates a fresh and deep dive report that will then be included in the overall, response that DeepSights produces, allowing you as the end user to automatically and seamlessly make use of it, but also to have deep insights then contrast and correlate that with other sources and other knowledge and research you have available into the final outcome and and product. And, of course, you can also, go back then and drill down, deep dive back into the Nextiva platform and use all the fantastic capabilities that Mario showed you that they support on their platform. So a very easy way to bring Insights to life, to integrate it into insights and also then to use the DeepSights’ capability of very easy and low barrier, low friction access to make it accessible, to everyone in the organization. With that, we’re concluding our look into the future, our look into foresights and trends. And, of course, if you’re interested, please do reach out, talk to us for a demo. And now we will turn to another situation, another scenario, another use case. And for this one, in a short construct from the House of Insights will be joining me now. And here, the question the background is, what if we now realize we have asked our question, we looked at all the data, we turned all the stones, and then we find out that there are still research questions we cannot answer as yet. So do we have to do research, or are there other ways? Sure. Oh, thanks, Olaf. My name is Shute Kollinstra. I’m a partner at the House of Insights. The House of Insights supports companies to upgrade the level of the insights discipline. We are supporting companies in pitches, for example, a new supplier for brand health tracking, which are often time consuming and an additional burden to the daily business. We’re also selecting and implementing insights engines for organization, guiding in the application of generative AI, data science, and marketing mix modeling, and how to organize your data household infrastructure. Market research is also considered as slow and expensive. Many companies require the insights departments to deliver cheaper, faster, and also generate more impact. Cheaper is more than self explaining. Faster and impact are closely connected. More impact means that it is better to be able to feed the business leaders on time with eighty percent of the required information than with ninety nine percent would take more time while in the meantime, the business decision has already been taken. Using data marketplaces, like data rate looks as a very attractive alternative. At dated marketplaces, millions of data sources are offered. However, consumer, insights managers are often not knowledgeable about these sources, and consumer insights managers might have doubt about the data quality. Too often, it happens that consumer data is only trusted if it is acquired according to the methodologies from the consumers insights managers. Karlsberg had several inner business issues, which we piloted, and I will treat one issue more in detail. The current economic crisis impacts consumers in purchasing behavior. Karlsberg wants to know if specific groups can be distinguished concerning the different consumer strategies, clustering based upon changes in consumption, spending of beer, but also within beer? And if explicit information about beer is lacking, could then possibly an analysis be done based upon similar products as beer? Consumer research insights case would be huge. Meaning, high sample with a complicated questionnaire and taking considerable time. In several countries, you have household panels data, which could be exploited. However, these are often too small for detailed switching behavior and focusing only at home consumption. Beer consumption has a significant presence in out of home, like bars and restaurants. The analysis will need some creativity and is not straightforward. It is not possible to automate the analysis. This means that finding appropriate data sources is the main goal. In the jungle of millions of available data sources, it is time consuming to find the suitable one at data marketplaces like, for example, data rates. In the case that we could automate the search for appropriate data sources that would save a tremendous amount of time, so could DeepSights support us. And DeepSights did a great job. It gave us two solutions, but also explains why these sources are suitable. The first solution insights of combining two different sources in which one contained the purchasing behavior and the other one, the social demographics. The other suggestion was a data source that collects data on SQL level transactions on personal level, which could be analyzed based upon movements in time. Our conclusion, which refers also to the three three other pilots we did, is the suggested data sources by DeepSights the DeepSights make sense. The suggested how to exploit, analyze the data makes sense. The suggested data sources were not known by Karlsberg, and the the output generated by insights doesn’t take more than five minutes. Because the analysis requires relativity, each data source has to be evaluated on suitability by a human being. But the speed of, find to find these data sources is tremendous. Yes. Thank thank you, Srid, for taking us through this little case study. And and as Srid said, we have also worked with another partner, part of DataRage, who are a commercial third party data marketplace, that, helped us in running this project and who we are also looking to integrate into the DeepSights platform so that what exactly what Shur has explained will again become part of the standard experience so that in addition to a report, for example, which compiles all the information you have already, it will also then come back and be able to give you recommendations for existing third party datasets, which may be usable and applicable to help answer the question you have and even, provide you with certain instruction and idea and direction of how to use and evaluate such datasets to make sense of them. So that is another very interesting angle to open up, the the view on what is available in terms of data, especially also for cases when the existing information is not sufficient. But at the same time, also opening the possibility of maybe not having to do bespoke research and rather reuse content and datasets that exist already in the market. As I said, we’re integrating it with DataRade. And again, if you’re interested, please do reach out and and talk to us for, preview access and also evaluating how it applies to your use cases. And now for the last, the next, the final section of our session today, we wanna talk about having done research done, having survey response data available, coming back from the research, what can AI do and how can it help with that? And for this, I’m John Berglion from Inspirience to talk about, what DevConnect technology can do and, of course, in turn, again at the end, how we can all bring it to life also then within the context of DeepSights. Guillaume, welcome. Hi, everyone. My name is Guillaume Aemeti. I’m cofounder of Inspirient along with doctor Georg Wittenberg. Inspirient, we founded in twenty sixteen to essentially save us, from a lot of hours, analyzing, data whilst we were management consultants. So I was at Deloitte. Georg was at the Boston Consulting Group. And, essentially, we found that we were spending we thought we were spending too much time trying to find insights in in data, in Excel, pivoting in all different ways, and not enough time kind of thinking about the results and what they could mean to our clients. We both have PhDs in computer science, so we thought, well, surely there’s a more intelligent way of doing this with machines. So, in twenty sixteen, we started building the platform for analyzing structured data, quantitative data with a machine to mine it for those insights and give us a prioritized list of the the insights that we should be, thinking about. And so what what is the problem that that we’re solving? So, not only do we analyze, general, structured datasets, tabular data. During the pandemic, we partnered with, Kantar Public, now known as Varian, to help them analyze, survey data, quantitative survey data. So their problem was that they were taking weeks to carry out their, survey quality assessments, to do all their, crosstabulations, to then run statistics on these crosstabulations, and then, create their charts for, and reports for sending to their clients. So they they came to us asking, well, you analyze tabular data. Do you analyze survey data? And it turned out that we we were analyzing the survey data, but we were missing these kind of specialist, tasks that they were doing. So we spent some time with their senior researchers, and we added those tasks that they were doing that were missing from our platform. So we’re now doing all the survey quality checks that, they would be doing straightliner analysis, speeders. And we also added all the specific stat testing that they would be carrying out and looking for, the kind of patterns, relevant patterns that they would be looking for from the crosstabulations. And so the the the the solution that we have is, depicted here in this diagram. So we’re taking raw survey data, respondent level data, which can be in Excel, SPSS, CSV, even SQL data. Our platform will mine that data for the relevant patterns, and it will actually generate by itself automatically, the the the the deliverables that you would typically get from an agency. So as you see on the screenshots here, with the the platform is generating slides with natural language text describing the chart and the highlighting the key pattern in there. And so you can now search for, findings, much more easier than having to dig through it yourself manually through the data. And so included in that is, the consolidation of multiple data sources it can be. We can analyze one or more data files. We’re carrying out that quality assurance, so all those data checks that senior researchers would be carrying out, the, statistical testing, and, trying to find those significant patterns and generating all the deliverables. So that process that used to take their researchers a couple of weeks is now done in, minutes or or hours for larger surveys. So as a as a actual client example, Alexandra is one of those senior researchers at at Varian. She has eleven years experience there, and she was taking multiple weeks doing all this. Also her skill set is is is rather, large. So the the things that she would be doing sometimes even more advanced, like driver analysis, multivariate logistic regression analysis. And what they she was having a problem with before in Spoon was, having to decide which questions to focus on because sometimes they’re carrying out rather large surveys, and they couldn’t analyze all of the questions because they just didn’t have time. So they would focus on ones that the client cared about, but then they would be leaving value on the table because they couldn’t do the analysis on on all the questions. So now within Experian, they’re able to do those analyses, within, minutes to hours. They are, able to junior researchers are able to use those methodologies that were previously only available to the, more senior researchers, and we can apply this at scale. So they don’t they no longer have to focus on only the key questions, but now they can analyze all of the questions and make sure that they’re not, missing any key insights. So the the the new process now with Inspirion is essentially a four step process. So the first step is to, submit the the the data to the platform. So there’s a a data upload page where they can upload, their data files. Once uploaded, the platform will do an initial scan of the data and try and classify the variables into dependent independent variables, the demographics, variables, the meta information, also try and identify those variables that help us do the quality assessment, like survey duration. And then what we can do is review that classification in an optional guidance step and correct it if it’s wrong. And that correct that correction step is also then training the system to understand it the next time it runs the the similar dates that was the the same variables. So when when we start the the the analysis in full, it can take anywhere from a few minutes to a couple of hours depending on the size of the survey. And during that time, it’s doing all the analyses. It’s looking for those insights and generating all of the output, that one would expect. And so the third step of the the process is to review that output. We’re generating output as PowerPoints. We’re also generating those Excel files or the cross tabulation tables that you’re used to, and, some additional files as well. The fourth step is to essentially, now that we have all of our data analyzed, we have all of these insights, how do we find them? And so the last step is to be able to discover them. And for that, we have a user interface where you can, a little bit like Google, be recommended, certain insights, and compile your own reports. But how does that fit within DeepSights? Well, so DeepSights is actually the perfect platform for helping you find and search for for those insights. So the way that we’re integrating within DeepSites is essentially Insights is doesn’t have to have a user interface to upload data. We can actually connect via API to a data repository with multiple, different data files in there and just allow Inspirion to mine them and generate all of those outputs automatically that it would generate. And then we can connect that to deep insights. And then deep sites can be used to retrieve or search for the those insights. And at this point, I’ll hand over to Olaf. Yeah. Thank you, Guillaume. And, of course, you may guess the pattern by now. We then, on the DeepSights side, can again also speak to, if you will, the experience AI, can make use of those findings that have been identified and can then, on the fly, generate new, well, insights based on the patterns, based on the findings, in response to questions that people ask. For example, then describing, demand occasion and linking that back to the underlying research significant research findings, again, cross referencing the underlying findings on the Experian platform so that, as always, there’s full transparency of in terms of where does all the data come from, and then there’s the capability to drill deeper and learn more. And this is, of course, very, very powerful because it enables you to use the information that you have paid for already to further mine it for new business questions, to further learn from the data that you have already, to answer or as as Leon said, to leave no value on the table, and to also then even begin to correlate that at this individual panel level across different sources and across different studies, and you can potentially learn more by connecting those dots. And again, also here, this is something we’re working on with Inspirent at the moment. And, if you are interested also for this piece of a partner integration, please approach us, and we’ll be very happy to to work with you on on a preview access to review it. And with that, we are through our, four segments, and I’ll I’ll hand it over back to to Kimberly. Yeah. Thank you so much to all our speakers. I’ll ask you guys to just come back on stage because now we are going to enter the Q and A portion. So if you haven’t gotten your questions in yet, please do use the question tab in the bottom right hand corner. However before we start things off, I actually have a question for the audience. Obviously, all of us here are big proponents of AI. We are the ones developing the tools but we wanna know that how you guys as the end users are currently using AI. So whether this is something that is currently part of your daily workflow, whether this is something that you are planning to, integrate into your daily workflows, just haven’t had the time, or if you are more on the skeptical side and need, more information. So just out of curiosity if you guys wanna go ahead and respond to the poll on your screen. This you know is something that we’ve found maybe people in the insights and and research industry are slightly more skeptical than, people in other industries when it comes to AI. And, we are tracking how that is changing. So, looks like we have a a very, enthusiastic crowd here, though, with, nearly sixty five percent already using it. So we we are the early adopters that Mario loves to see in his work. Alright. Over to the questions. So I I know some of you have specific questions about integrations as it, applies to your own specific platform so we’ll answer those separately. But let’s start with a broad question. Maybe back to Olaf to begin. How do you deal with general data privacy mechanisms? Obviously we’re dealing with a lot of information uploaded to a system. This information might not be while we while we are proponents of accessibility maybe not everyone within an organization should or, has the, should have the ability to access it. So, what mechanisms are built into DeepSights to afford enterprises this level of data privacy? Yeah, that’s an excellent question and one that of course we hear quite often. Of course fully understood that they are, especially in larger organizations, is a need to segregate and to ensure also, access controls and visibility. So, our platform does support that, in the net effect being that that we can assign documents, for example, to access control rules, and that in turn means then because users who would approach DeepSights to obtain answers or to get reports would only see that part of the information universe that they’re entitled to see, and and their answers and reports would really focus on that body of content, to make sure there is no leakage, if you will, of information. And, also, of course, everybody can rest assured the data, that any of our customers contributes to their insights instance is, of course, private, to that customer and not shared or used in any other way, also not shared or used for training, in AI. Yeah. I think that is something that, a lot of these platforms share that, you know, when we say it’s trained on your data that that training does not bend weak over to someone else’s, system. All right. Here’s one for Experian. So Guillaume, you’re up next. If you have conducted multiple surveys, where will Insights question be applied to all surveys or does the platform make a decision which survey to draw data from? So can you combine the results from different surveys that cover the same topic? Yeah. So so the idea of Insights is that it’s monitoring, a repository with, data sources. So when as soon as a new survey has been carried out and the data exists in that repository, it will be analyzed. And then the findings from, the the survey will exist. So as soon as you go into DeepSights and ask your question, it will have access to all of those findings and will retrieve the right ones. So it’s it’s not the other way around where you ask a question on DeepSights and then insights we’ll do the analysis. Yeah. It’s it’s the other way around. So, yes, it will analyze all data added to the repository. Great. And this is one for Mario. I know in your presentation, you highlighted some of the the big social media platforms that you you currently draw from. Obviously, we have, some people in the audience that work for global enterprises that might want to have, trend forecasting for, countries such as China that that use social media platforms, beyond, you know, Twitter, Facebook, etcetera. So, what can you say about, the the social trend forecasting offered for those regions? Yeah. This is, something that we do outside of, the Nextatlas platform. So, why is Nextatlas, is actually limited to Instagram x, YouTube, and and so on and has a global coverage, for the majority of the markets that use the, let’s say, Western social media networks. Whenever, we need to analyze specific key markets, like China and in the past, Russia, we go to look for specific and we enter specific social media networks of those markets. But this is not something that we do, within the Nextatlas platform. So it is something that becomes as a bespoke survey or a bespoke analysis that, can be provided. And therefore, we need to define or understand specifically the brief and how we can, actually, do it and do it together with market logic. Got it. Yeah. And I know that, you know, this we we fit a lot of, explanation into a shorter period of time here. So I think all of the speakers could say that there’s a lot more that the the different companies offered that we didn’t share today. So, again, I just suggest everyone get in contact with us if you want to have preview access to these integrations, or, just find out more about how they could work for for your specific use case. If anyone has any questions, just, get them in now. Otherwise, I will, start the conclusion because, you know, we’d like to wrap this up on time. So thank you everyone, again for for sharing. And short, I know you were speaking also on behalf of DataRade, so thank you for taking that that up as well. We can, share more about the DataRade partnership for those interested. Insights. So that concludes our event for today but that does not conclude our events in general. If you are interested in this conversation and want another, panel discussion that focuses more on the evolving role of insights teams, happy to say that we’ll be joining some speakers from Estee Lauder, FGX, Cary, and Fuel Cycle next week. This is a conference hosted by TMRE, our longtime partners, but registration is free. If you are interested more in the technical side of things, we are gonna be doing a longer demo of DeepSights and talking to other companies that are, putting AI powered insights platforms and related solutions out into the market. And that is with Isamar, which you guys are probably familiar with. Otherwise, just give a wave finally, Mario, Olaf, Sjoerd, and Guillaume. Thank you all for coming. It was a very interesting session. We will be sending out this recording and if you have any questions, you can get up, in contact with me personally, and I will make sure those questions get to the right person. So thank you everyone for joining. We hope to see you again soon. And, yeah. Goodbye. Goodbye. Same next to you. Bye. Bye. Thank you. Bye. Thanks, everyone. Bye.