WEBINAR: Transforming Market Research Workflows with AI

Okay. Hello everyone. Welcome to today’s webinar Transforming Market Research Workflows with AI. My name is Caroline and I’m just here today to share the agenda, do a bit of housekeeping, and then moderate the Q and A at the end. So as people trickle in, let’s go over the housekeeping first. We will be joined by two speakers today, followed by a Q and A at the end. I’d just like to direct you guys to the questions tab in the bottom right hand of your screen. You can submit a question at any time to any of the speakers or even both but I just ask you that if you want a specific person to answer your question please just include that in the question and it will make the q and a go smoother. Other than that, this session will be recorded. If you have to drop out a little bit early, we will deliver it fresh to your inbox tomorrow morning. So that is all good. And if you have any, more logistical related questions or questions that we can’t cover in the q and a session, we will follow-up offline via email with you. So that’s that. Onto our speakers. You can see them on screen now, but I am happy to have with us MarketLogic’s own, Olaf Lenzmann. He is serving as our chief innovation and product officer and is the driving force behind a lot of the innovations within our DeepSights solution, that you’ll be learning about in a little bit. He is joined today by our friend, Louise Hitchen from MMR Research. Louise serves as the head of head of digital research solutions at MMR, leading the internal implementations of new platforms and tools, ensuring the everyday integration and, usability in workflows. And she is particularly interested in harnessing the power of GenAI, for a positive impact on the market research, community. So she is a perfect spokesperson to join us here today. As for the agenda pretty straightforward. Olaf will start things off just to give us an introduction of what AI for Insights is. We’ll be generally discussing what it looks like when deployed for the market research and insights function. We all know it’s an exciting technology, but does it stand up to the data quality needs, the regulatory requirements, etcetera, that our insights teams, work towards. Then we’ll hand it over to Louise who will give a look at how MMR has gone about evaluating and integrating, DeepSights and Gen AI solutions for their own purposes, and then Olaf will come back in for a little bit of a future outlook. So, yeah, we’re hoping to just explain a bit of the what, why, and how when it comes to using this technology in our industry, and, we hope to foster a pretty engaging discussion. So get your questions in. And, without further ado, I’ll hand things off to Olaf. So thank you guys for joining. Yes. Thank you very much Caroline and Hello, and welcome from my end. I wanna talk about the first segment, AI for insights. What is that? What do we do there? And, of course, I’d like to start off with a very tiny little bit of background about us, MarketLogic. Who are we? We are a provider of a software service solution for our customers, which enables them to bring together all their market research intelligence, all the knowledge about consumers, brands, markets, competitors. And with the help of AI, that platform make the best use of that information to bring the insights to life in the operations of the business to make winning decisions ultimately. We’ve been in the market for a long time, and we have a roster of, wonderful customers we work with. And we also do have a large, list, a long list of partners that we integrate with so as to make sure that all the data comes onto the platform, not only the data that our customers bring, but also the data that is either public or that is maybe syndicated or paid is on that platform. And while you see that we’ve been in the market for quite some time, AI, of course, has always been an important element. But as we all know, there has been a major shift in the AI capabilities recently in our generative AI journey, specifically with those recent, technologies in terms of large language models, which enable fantastic new use cases that are also very important and exciting in in our domain. That is a journey we have started already late in twenty twenty one when the original GPT three series of models came out, and we immediately thought this is a great technology to tackle some of the challenges that we see in our customer base when it comes to bringing the insights to life and then bringing them to where the need for the information and where the decision is. And so we started out working on that with our lead customers. And since for more more than a year now, we have a product in the market also, which we call DeepSights, which has this Gen AI based, core. And we have also been and continue to be constantly innovating with further capabilities around DeepSights, like we introduced last year, the ability to generate full HAWK reports that integrate the knowledge across your knowledge base. And now we also have recently launched, our APIs to integrate that capability into other business processes and tools. The question, of course, then also is what sets us apart with our solution, also apart from other maybe AI driven, Gen AI focused, approaches to deal with data, to work with market research and insights data. And we like to think that, first of all, a key differentiation is that we have tailored and trained the AI that we use specifically for this use case of insights and market research. And I’ll speak a little bit more about that, in a moment. But that really makes a huge difference when it comes to the nuances and the specific aspects of how to deal with the data and the information and how to interpret it, that we encounter and that we work with. Secondly, we also bring the insights data ecosystem, as we like to call it. So we bring all those preintegrations with external third parties. But, of course, we also bring the connectors and capabilities to integrate all the information that our customers have in house and to onboard them to the platform so that the platform has a three sixty degree view on all the content and knows everything, our customer knows. Then we do focus very, very much on enabling, as we say, next generation I go to market professionals, and we’ll touch on that a little bit more. But it is mainly around our abilities to also bring our information with the help of AI into other processes and tools so to embed ourselves both to serve the end users with the information that we have, but also to serve and integrate with other processes and systems. And lastly, also very importantly, our system, our platform is, of course, secure, secure, and robust to the level that our enterprise customers require. But also it is adaptable in the sense of being able to be tuned and tailored to the best practices you have. So whatever the best practices of a company are in terms of dealing with insights, they can also be configured, if you will, into the AI for it to take it into account. And, therefore, not only be, especially trained AI for insights, but also your AI, for the US practices. And a few words on what is so special about the training and the way we approach AI for market research. Well, first of all, at the center, at the core of things, of course, we ground everything in the knowledge and insights of our customers. So we take all that big corpus of knowledge, and we, with the help of AI, extract the findings as we like to call it. So go through all the materials, peel out what are the relevant and reusable pieces of information, not only peel that out from textual content, but also now use more vision models to, extract information from visual presentations, from charts, also understand with the help of the AI, what is the context of the research, when was it done, how was it done, what market does it apply to, etcetera. And then when we’re trying to support specific use cases, we, of course, use tools like modern AI driven semantic search to find relevant content in the knowledge base. But not only that, we also have proprietary models, that we’ve tuned to help really narrow down and pinpoint which pieces of evidence that we found in the corpus really do answer the question exactly as it was asked and really qualify to be reliable and trustworthy evidence for generating the answers and generating reports. So with the help of these techniques, we’re able to absolutely minimize hallucinations that are a very common issue in this, large language model space. And, also, we’re able to make the system really understand and honor the nuances between what is the exact market that you had a question about, how relevant is it, if the information is already older compared to other kinds of information, etcetera. So we need to take these, specifics of our domain into account. And with this underlying capability, we’re able now to help our, customers, stakeholders of our customers to use a very, very simple user experience, which we are all familiar with by now, when working with these kinds of AI assistants. With our DeepSights AI assistant, we let them ask a question, a business question, a very natural language, and within a matter of seconds, get the answer from the AI from across the entirety of the knowledge base focused on the most recent and relevant pieces of information that speak to this question. And, also, optionally, then ask the AI to go deeper and come back with the full synthesized report that spends multiple pages integrating, contrasting, and correlating all the information that’s relevant that was found. And these use cases are very, very simple and, very easy, for stakeholders, for researchers, for users to work with insights and at the same time also with some evaluations we have with some customers. We find that, the sheer efficiency gain in terms of the speed with which you obtain an answer that might have to take you hours to read and then hunt down, information is now solved in a matter of seconds. And, also, lastly, very importantly, this kind of very easy question answering interaction lends itself to bringing the insights to where the people are that need the insights. So we integrate this with tools like Teams or, Slack or Google chat, for example, or email So that it’s very easy for somebody who has a need for insights to get the answer without leaving their environment, without having to change the context, but doing it right there where they are. So this is what AI can do without any human intervention. But, of course, the human element is extremely important, and that’s why we also combine these AI capabilities with tools, for the experts, for the market researchers that can then use their intelligence and expertise and experience, to, for example, curate the most relevant insights on certain strategic topics that are, important to toward the organization and then to send that, information to disseminate it in a targeted fashion to exactly those stakeholders in the company who are in need of that information. Or similarly, to also, provide those stakeholders with up to date AI driven newsletters, news reports that, again, are tailored exactly to their profiles, jobs, and context. And, also, to use our platform to help drive governance and collaboration when new research needs to be done, when there is a gap in the knowledge base, and here also to use best practices, templates, approvals, and direct collaboration with suppliers to make sure that filling those knowledge gaps through market research projects, is efficiently done. It’s done according to best practices, and it perfectly also dovetails into existing knowledge repository. And lastly, as I mentioned already earlier, we connect here, the entire ecosystem. So other, data sources and systems in the organization inside and also outside with third parties. Lastly, and one very important ingredient we believe to enable this, AI driven, workflows, we also provide a rich set of APIs that enables our customers to integrate these capabilities into other business applications that they may have where they, again, also want to expose insights right in this environment to their users or, similarly, into automated business processes that they may have where they, again, through APIs, can retrieve contextually relevant and trustworthy insights and information that then can be passed on to further downstream steps. One example of this for you are, also, applications where we then feed our insights into other specialized AIs downstream applications, like, for example, customers we have who have developed concept generator AIs where then our, inputs serve to brief this other AI, this downstream AI, if you will, to exactly tailor the generated concepts, for example, to all the insights about the consumers who are targeted. And finally, we also can complement generic AI capabilities, generic AI assistance, and agents like Microsoft Copilot, which, of course, is very powerful, omnipresent, and well integrated assistance. But we, here in this case, can integrate ourselves to enable the ecosystem to delegate the insights specific tasks to our insights, AI and, therefore, to leverage all these optimizations and all the trustworthiness and all the data integrations around that space while still using Copilot or similar offerings as the overarching umbrella with which you would interact. So there’s a variety of ways of how now the capabilities that are, yeah, I’ve driven these days, can be used, to leverage the information, pull it together, aggregate it from all the different angles inside and outside the organization, make sense of it, enrich it with the human expertise, and then play it out either to end users, into other tools, into other processes, or to complement other AI systems. Many channels and many opportunities to bring insights alive today through these technologies. And that’s a quick overview of what we do. And now I want to, hand it over to, Louise to talk us through through her case study. Thanks, Elas. Hi, everyone. So first, I’ll just give a little bit of, an intro, to MMR, and myself. So MMR was founded about thirty five years ago, really with the idea of bringing together consumer and sensory research. So since then, the MMR family has grown as a collective of of brands really working together, to support, clients with the right skill set at any stage of the MPD cycle. So we work with some of the biggest FMCG and and CPG brands to to really champion the the product experience. And MMR alone have about five hundred people working across nine different offices, and and we’ve grown quite a lot recently. And it was really this rapid growth that was a key factor in us choosing to explore a partnership with Olaf, and his team, which I’ll come on to a little bit more in a minute. But what I also wanted to to talk about was we’re we’re in quite a unique position and quite fortunate within MMR to have a very, strong support from our SLT for internal innovation. So we do have an in house innovation team called Nova. And although we still believe, as I left mentioned very much, that the human is at the heart of everything we do, but Nova, a team within MMR who are dedicated to exploring, experimenting with different technology, really to enhance the not only quality of insights that we can generate for our clients, but also how how we work, as a business. So the platform agnostic goes through. You can see the the kind of rigorous, evaluation process here to ensure that anything that we do deploy is of of tangible value. So not just tech for for tech’s sake, but it that it really has an impact. My team then within MMR, as as Kelly said at the beginning, I lead on the implementation and usage of these different platforms within the business so that we can support adoption, make sure we’re fully embedding, these tools, and making sure that we get the most out of them. So as I mentioned, alongside our client research needs, Nova also look for opportunities to improve our own internal ways of working. And this is really why we started to explore a GenAI knowledge solution. So we were seeing how technology could help us facilitate the flow of information within MMR in in a more effective way. So what were we looking for? We were really looking for accessibility. So we wanted all teams across the business to be able to expand their knowledge. Secondly, we wanted an efficient solution, to free up resource that was spent going back and forth, fielding the same questions. Thirdly, we wanted something quick, so as near instant as possible. We didn’t want people to have to wait for for the answers that they were looking for. And, firstly, we wanted to ensure relevancy. So particularly, as I mentioned, with the combination of our team growth and our ever evolving toolkit of methodologies, we re really wanted that one source of of truth and to make sure that anything, that that was returned as a search was was really relevant. And linked to that, again, as Olaf mentioned, accuracy and and surfacing that unbiased information was was really, really key to us. So although our general interest in a, a GenAI solution, particularly for knowledge management, was internal to start off with, We wanted to keep an eye open knowing that it’s a a rapidly evolving space. We wanted to keep, open to future opportunity spaces and make sure that really any partner that we we chose could move with us. And that was again one of the, reasons for for choosing to partner with with Olaf and his team and and deploy DeepSights. So, yeah, why did we choose to to partner with with OS? So I think we’d we’d gone through quite, a a rigorous process, as I mentioned, with with the Nova team and the evaluation, but I think the importance of, you know, protecting the the the data that we’ve got, that that privacy, the security of our information, was really paramount. And from a compliance perspective, yeah, DeepSights definitely checks, the box. And I think the second is integration. So some of the other partners we looked at, wanted to migrate everything, but we really needed a solution that fit in with our own workflows. I’ve mentioned again how how important that is, that the seamlessness of a a connection with existing workflows, connecting with the apps that we already use and the the tools that people are already familiar with. Not only from my perspective in terms of adoption does that help, but also, it just just means that it’s a more, seamless approach when when we’re trying to integrate it, yeah, in into our workflows. And thirdly, I think the focus on the insights industry does mean that the the models are trained on the data that we’re generating, and and the importance of that is, yeah, really, really crucial when you’re looking at the accuracy information and and the the searches and the the results that are being returned. And they understand our needs as a research business, which is a partner going forward, is is really crucial to us. And I think in terms of how DeepSights has really impacted our our workflows, we are still in the early stages. And, as as I’ve mentioned, DeepSights is continuing to evolve and and grow and add add new features, but we’re seeing some really positive, initial feedback from from our team. So the way that we deployed it is internally at the moment across our our our toolkit. And people across the business are starting to see DeepSights as a a really useful team member, so someone that can quickly summarize a lot of content, but really importantly, being able to cite its sources to pinpoint for for further human investigation. And in terms of mechanics, I think the the way that we’re seeing, a lot of uptake in the business is is through the team’s integration. So I mentioned in the selection criteria that being able to integrate with our existing apps and tools was really important. And we know that this is how, our teams are used to interacting. We really wanted to mimic that that conversational behavior. And another way it’s being utilized is to really generate those thematic, thematically structured summaries, really, and give a bit more depth on on a particular topic that that people might be interested in or or that people might have questions on. And I wanted to share a couple of more specific, use cases where we’ve seen it, as a really powerful tool. So, Insights doesn’t sleep, I think, unlike, hopefully, some of our our colleagues. You you can get an instant answer really whenever you ask the question. Doesn’t matter what time it is. Insights is always available, to answer, and it facilitates access to a wider pool of inspiration. I think as I mentioned, sometimes you’re fielding the same questions or you’re asking the same questions of the same people, and it’s that same information circulating. Whereas DeepSights encourages that cross team learning and access to that that wider, knowledge base. And and I think in the same way, as we’re talking about democratizing access to to insights, where we found it really useful, as I mentioned, with the team growth, it’s really powerful for onboarding. So whether a full immersion, for a new starter or even a a role, switch. I I came back from maternity leave about six months ago, and and, when I came back, I found it really useful for reminding myself of a few things. We all know how many acronyms we have, in our various industries. So just refreshing my memory of of these things, and, there’s no such thing as a silly question, when you ask DeepSights. So no embarrassment with asking a colleague something that that, you know, you probably should have, remembered or particularly for junior team members if they, might be a a bit hesitant to ask something that that they’re unsure of. We found that that’s been a a really, really powerful, tool to to democratize that that access to information. I think this is a a a quote from one of our team that sums it up, quite nicely and captures the the sentiment. So someone said that DeepSights is an AI powered search engine, on steroids. So, again, just shows the impact that it, yeah, has has made to to different people’s, experiences of, yeah, that that access to to knowledge management. And I also wanted to share a little bit more about what we’ve learned from deploying DeepSights and a little bit about, the enablement process. So I think, firstly, what what we found, pretty quickly was solid content foundation is is really key. So it’s highlighted some of our own, potentially inconsistencies within our own materials and and knowledge base that, was was really key for us to to address. And I think, secondly, we found that starting small was was, good for traction. So finding its usefulness really memorable, finding a key hook. Don’t try to be, everything to everyone at the at the start. Finding that that key, killer killer element that it can help people with was was really important in adoption. And I think they’d be linked to that with any behavior change. It does take work. That’s what I spend a lot of time doing, with with my team in terms of implementing different tools, but make it as easy as possible. Integrate, as I mentioned, into existing workflows wherever you can, and that, really, really helps with with uptake, and people may be able to jump into using it immediately. And finally, I think, yes, stay stay strong with the implementation of of new tools. In in my team, we often talk about having a consistent message, saying it’s right before it sticks and, not not giving up because, you know, people don’t see it as an instant hit. But I think take the feedback on board, continuously nudge, to towards that end goal. And we are we seeing the power that that can have, but definitely give guidance along the way. And as I mentioned, yeah, just just find those those killer, elements that can really help, make an impact on people’s day to day workflows, will be my learnings from from this deployment. And then in terms of a a future outlook for for where we see, the, kind of, evolution of of GenAI and and impacting our our workflows, I think the way that that we see the future of, market research in general is AI enabled. So I think we’re, as I mentioned, continuously leveraging new technology in incremental ways really to enhance our knowledge, and our creativity and fueling better connection with consumers. That’s really at the heart of of what we do. And I think the first pillar where we we see that, is synthesis. So I think, naturally, where GenAI is really exceptional potential as we’ve already seen internally with with DeepSights, we’ve seen the benefits of using AI to speed up desk research, tapping into that readily available data, for example, through social listening. And we started to see the benefits of AI enabled analysis for even our primary qualitative research. But as I mentioned, our our approach is still very much human centered, but we use AI and tech to enable those humans really to work more effectively. So, for example, removing language barriers with, some of the the more accurate, translation tools, making it easier to analyze content from different sources and and pull apart the the segment differences. And then I think the next pillar that that, we really see, within MMR is is, under the create pillar. So we can leverage AI to inspire and and build hypotheses at the start of a project, whether that’s through creative prompt engineering. For example, we work a lot with a lot of, obviously, food and beverage brands. So we might ask a a Gen AI, tool to give us suggestions for content inspired by a specific cuisine, for example. So what it delivers is, of course, not the full answer, but it can help us to push beyond, the obvious, giving us potentially strategic starting points to really fuel further conversations. And then I think we’re also seeing really effective use of AI for visualizations, so combining, different text or image tools. So MidJourney, DALL E three, things like that, particularly combining that with our own internal creative expertise. That’s where we’ve seen some really effective use even within, for example, live focus groups to visualize consumer feedback in the moment, in real time, or used as prompts, to help them, you know, consumers better articulate their thoughts or even in deliverables to bring to life consumer ideas. That’s where we’re seeing the the power of AI tools in visualization. And then I think the final one is deploying AI in in really, really powerful ways to better connect with consumers. We’ve got embedded capabilities within MMR for conversational AI and things like chatbots, both qualitatively and quantitatively. And what we’re doing now is embedding our IP into those models that we use to train the bots and make sense of the data. And I think developing specific solutions, within MMR, we’re building a a sensory bot, which we’re, training a lot of, data on a lot of models, and that will help us deconstruct the the sensory product journey from a consumer perspective. Again, getting to the the depth we need for at scale. So I think we see opportunities across all of these these four pillars, And I think, yeah, there there’s definitely really exciting work being done within insights space. So Olaf’s gonna talk a little bit more about, the the future, role of AI and and and workflows within insights site. So I’ll pass over to Olaf now. Yes. Thank you. Thank you very much, Louise, for sharing this. And, yeah, a few words from from my perspective on where we see the AI capabilities evolve specifically from our lens, from the DeepSights angle. One thing, of course, we continue to work on is further integrations, out of the box integrations with more partners, with more data sources. Data, obviously, is the fuel for the AI to connect and consolidate all those dots. So we’re working with a couple of, partners on integrating, for example, currently more video platforms or also other, paid content sources. So over time, to enrich further and further the the breadth of the information that the AI can draw on, with a clear aim of making all of this out of the box and minimizing any any efforts on the side of the customers, the users to bring the data in. Secondly, another very important direction for us is now to based on the foundational capability that we have, which is being able to take a business question, take problem, and find reliable, robust, and trustworthy information, speaking to that problem, and now to build further, analyses on top of that that follow now a more complex, brief, and objective, and that can then also with the researcher together in a in a joint human and artificial intelligence, journey be used to come up with new derived insights to assemble and author new information with the AI helping to uncover pieces of relevant, findings and making connections and proposing elements, but then having a tight, experience for the expert, to take that, compile it, curated, elevate it, and to produce new outputs. That is very strictly and strongly linked also to what we call your AIE five best practices. Of course, every expert team has their own, best practices. They have their way of working, and they have the established and proven ways of analyzing, synthesizing, information. And while we can already, give certain guidance to the AI today, for example, to, focus on certain sources of data for certain kinds of questions. Clearly, the the journey will go to further more complex guidance, more complex practice practices. For example, setting up preconfigured best practice analysis elements that are in court in accordance, with the way of how a customer wants to do a certain kind of analysis. And then being able to run these with the AI and use that again in this human AI, combined journey, to come to more, relevant results and insights at the end. And lastly, what also was mentioned by Louise, the creative field, of course, is very strong for AI, for generative AI. We have some customers already that, link to our systems to take insights from us to to guide the creative process, help guide the creative professionals on the next step, also here, on the one hand, for sure, we see a lot of potential, and interest to deepen this relationship. But, also, this can go the other way, and it can mean, taking a creative proposal and then critiquing it against what we know from our insights, from the consumer, from what has worked in the past. So it is not only fueling the credit generation process, but it can also be, yeah, playing back insights against the concrete, idea and helping to iterate on it. So these are some of the main areas that that we’re currently working on and that we see insights involving. And, of course, this is a journey, and this will remain to be a journey for a little while. But there is already a very strong foundation, and I I hope, case also made that clear. And, of course, it is also a journey to bring such a solution to life, deploy it, and then bring it into an organization. And, of course, we are here and we’re very happy to also help you in any such evaluations. For example, with a free trial of our products to see how that fits and how ultimately for your specific use case, it can be proven to work. Yeah. And that’s what I wanted to share. And then I guess we go on to q and a. Yeah. Yeah. Let’s, let’s move on to the q and a. I have a few questions for each of you. Obviously, Olaf, a few more product related questions, and then, Louise, some more operations questions. If anyone wants to get in a few more, questions for this section, you still have time. Let me start with you, Olaf. This question comes from Steven. Are you training your models for each client you have, or are you training it on entire corpus of data that you have? I think this is especially relevant for an agency model like MMR. Yeah. That’s an excellent question. We do not train, for each and every customer, And we we have trained elements of the AI, but also importantly, we didn’t train the AI on to learn the research facts and findings themselves. We trained it to be a good researcher, if you will. We trained it to be able to work with market research and related information and to, if you will, look out for what is important in this context. And that that is something we have done with, pilot customer data, but that applies and generalizes very well across different industries and, you know, use cases. Yeah. I think, actually, it might lead into a question for Louise because, obviously, I think both of you guys mentioned that accuracy, verifiability, trustworthiness, these are, I mean, they should be top priorities for any AI tool, but I think especially for market researchers, we tend to value, you know, accuracy above everything else. And as a result of that, compared to some other industries or some other functions, there has been some hesitancy to go out and adopt AI tools amongst market research companies. I’m wondering, Louise, if you faced any hesitancy, in any of the tools or kind of solutions you you’ve set out to deploy with Nova, and how you handle, any resistance or kinda lack of trust in AI at this at this stage. Yeah. Definitely. And I think that there’s there tends to be two, I think, hesitations. One’s around the security of the data, and I think, that’s almost a hygiene factor with whoever we choose to to partner with that that we’ve we’ve kind of protected the data in that way. And I think the second is the the accuracy. The the tools that we use in particular with DeepSights, it’s all evidence, so you can link back to the the source data. And I think for, particularly our internal teams and our clients, I think that gives a lot of reassurance that, as I mentioned, we still got the human in the loop with any analysis that we do with any, results that it returns. So it’s still down to the human to, you know, interpret the the summaries that are produced and and tell the story as as it were from, those summaries. But the fact that we can link back to the original source evidence for us is really what, gives a lot of a lot of reassurance to to people. Yeah. And also just starting small, getting one team used to it, and rolling out from there. Yeah. A few more questions. I think this is, for Olaf from Elena. Can Insights integrate consumer research data with the consumer panel data, e g trial repeat of certain brands or SKUs, and retail panel panels slash POS data, e g market shares, distribution, etcetera, to create a fuller story? Yes. Interesting question. Of course, on the one hand, you may have, top line reports, etcetera, from these sources, which are one path of getting into the AI. And there is another path which revolves around, a partner we have and work with together where we can, start on the individual response or transaction level data to identify patterns with the AI. And then based on that, answer questions, that you may have. So, it depends a little bit on the specific setup. There are a variety of ways. I have to follow-up maybe to illustrate a concrete scenario that you may have in mind. Yep. Maybe something a bit more general than, can this AI interact and talk to proprietary AI avatars? I think you spoke briefly about this with the API slide. But Proprietary AI avatars. So we, at this point, don’t have a voice or avatar interface. We can, of course, through APIs, we could be driving such an interface. It depends a little bit on how, again, you would envision using that and bringing it to life. We do have some internal experimental voice interfaces, but that’s not part of the product at this point. That’s also always an interesting feedback from the potential user base. Are you interested in a professional context to talk to the system that has its pros and cons, but maybe, it’s also a little bit awkward depending on situation? So long story short, it’s not necessarily an integral part of product, but we do have ways to interact with these things or to feed them if there’s a case to. Maybe in the future, it’ll be a living, breathing robot from DeepSights, but at that time, it just text. Alright. Can you describe how you protect proprietary IP if it’s using a third party GPT? So Well, all the IP of all our customers data, is residing in segregated places. So it’s it is clearly assigned to separate stores, and we do not train or fine tune the the model, the large language model on the actual customer content. We have fine tuned our system to to be able to, in general, answer questions based on market research in a in an intelligent way, watching out for the pitfalls that there are. But, as we as we, deploy customers’ information, that is processed by AI. Well, in this case, these are, for example, GPT models on Microsoft Azure with all the recording security compliance and data residency regulations. Also, none of the data is used neither by us nor by OpenAI or Microsoft for training AIs. So we do have contractual arrangements in place, obviously, that ensure that this is the case. And every customer has their own instance fully segregated and isolated from the customer’s contact. Great. I think we are cut we’ve covered most if not all. There are a few questions that have, you know, one or two word answers that I’m just gonna get back to, in my own time. But why don’t we end on a question for both of you, which is another sort of future outlook question, which is how do you expect AI to change the responsibilities of, researchers and insights managers? So maybe I’ll start with Louise, and then we’ll toggle back to Olaf. Yeah. I think it’s an interesting, one. I think at MML, what we’re really focused on is helping our clients connect better with consumers, bring them to life. And I think AI is enabling us to do that both in the way that we collect data with conversational AI, but also how we then can more quickly analyze the data. And then our job becomes more about the storytelling and the impact that we can have on clients’ businesses. So the way that we see it is it enables us to handle a large volume of of content and and data and enable us to focus more on, that that delivery to clients versus it being on the the management of of that data. So it allows us more time to focus on the bits that we hope have more impact with with clients. And Olaf, do you agree? Yes. Of course. I agree. I think there are several interesting dimensions. One is the efficiency, and also the ability then to focus on the higher, more value adding, elements of the of the value chain for for all the humans involved. Also, I do think that, this will enable researchers, to scale some of their expertise better by also imprinting some of their knowledge and and, best practices into the AI, which can then repeat that and and bring it out into all the different interactions when maybe today no researcher can be part of the discussion simply for scalability. That’s so only so many hours and the day reasons. So, in that sense, I think there will also emerge a certain discipline of how to, yeah, establish a governance from the technical side that, makes sure that the best practices and expertise is that the market researchers and insights teams have developed and that they own, so to speak, that they also come to life in this new digital infrastructure. Great. Thank you. Yeah. I think that’s the general consensus moving more to a strategic role, which I think, most people in the insights function would welcome. So I think both of you are doing, your part in in helping, the insights function transition to that that more powerful role in the organization. So thank you both for sharing your insights today. As I said, if if there wasn’t a question that was answered on stage, we’ll just shoot you a quick email, and we will be in touch with the recording. If you want to watch it again, or share it around, that will be available to you tomorrow. So I’m letting you guys go a few minutes early, but I think we covered a lot here, nonetheless. So thank you again for joining, and, we hope to see you at a future event. Once more, Olaf and Louise, thank you for for joining us, and see you again soon. Thank you all. Thanks, Callie. Thank you, and Louise for being with us. Bye bye.