How To Build AI Governance Frameworks to Scale GenAI Adoption

Two-thirds of consumer goods companies are leveraging AI, according to CGT’s 2024 Sales & Marketing Report, but challenges around governance and data foundations are keeping manufacturers from scaling these efforts. During CGT’s latest webinar, panelists from Microsoft shared insights about AI as co-pilot tools, and the necessary prep work CPGs must account for to build AI data governance and frameworks before they can truly embrace the technology and reap the rewards. From Unilever using AI-driven molecular chemistry simulations to Instacart enhancing customer engagement through conversational commerce and CarMax automating product descriptions, get real-world examples of the transformative potential of generative AI.Liz Dominguez: Hello everyone and welcome to “Navigating AI Adoption for 2025: Balancing Ambition with Responsibility.” My name is Liz Dominguez and I'm CGT's Managing Editor. I'm excited to welcome you here today.CGT publishes the Sales & Marketing report every year, and this year's report focused on learnings about leaders' plans for tech investments, how KPIs are changing, and top-of-mind consumer behaviors that will dictate the industry's path forward. Sales and marketing is under pressure. Consumer goods companies are looking at budgets a little bit more closely to counter factors such as rising inflation and lower margins. While leaders are eager to bolster their technology, their hands might be tied as stakeholders call for more detailed ROI assessments and budget oversight. AI specifically has posed a challenge in this space. While two-thirds of consumer goods companies are using artificial intelligence and machine learning, most companies remain in the exploratory stage, particularly when it comes to generative AI.So let's talk a little bit about how companies are walking the fine line between ambitious innovation and mindful responsibility when it comes to AI. So I'm excited to welcome to the virtual stage Dina Zhou, retail and consumer goods, Microsoft; Ranti Okunoren, principal corporate counsel at Microsoft; and Nydia Galarza Rosario, assistant general counsel at Microsoft. They're going to take us through some strategies to help consumer goods companies shape their 2025 AI frameworks. So Dina, welcome. I'm going to let you take it from here.Dina Zhou: Sounds great, thank you so much Liz, and it's really truly a pleasure and an honor to be with you today. Along with my colleagues, we will be sharing some of our observations by working with some of the largest consumer goods and retail customers to bring some insights to you about what we're learning, what we're hearing, and Microsoft's recommendation in terms of how to take generative AI and your AI ambitions forward.And so before we essentially get into what is the meaning of all of this, I just want to set the stage with what is generative AI and why is it so important? So the generative AI era is here. It took us three months to get to 100 million users on ChatGPT. And if we put this into context, it took 30 months for Instagram and 70 months for Uber, which are all very much part of our everyday lives now.If 2023 was the year that the world discovered generative AI, 2024 is the year organizations are truly beginning using and really deriving value from this new technology. In the latest McKinsey Global Survey on AI, 65% of the respondents report that the organizations are regularly using GenAI, nearly double the percentage from the previous survey just 10 months ago. As Liz mentioned, from the Sales and Marketing survey that was conducted and released in September of this year, 65% of the respondents expect generative AI to have a significant impact on the industry in the next 12 to 18 months. And so taking that in, and of course as part of Microsoft, that is part of what we're constantly hearing about and what we're really talking to our customers about as well.So I want to do just a brief primer on how we think about the era of AI. And the first thing to understand is AI is not new. In fact, some of the earliest work on AI was done in the 1950s and 1960s when researchers first began to explore the possibilities of creating machines that could mathematically reason like humans. They called them expert systems. Then we moved on to machine learning, which is good at finding and identifying patterns, but it wasn't very good at making decisions, but there was quite a step forward. And so fast-forward to the present where generative AI has sparked a wave of innovation as machines consume vast information in vast quantities, approaching what we know as general intelligence, which refers to a machine being smarter than any human being on a topic — take spelling bee as an example.So to put it simply, maybe when you're thinking about the term generative AI, think of something where there is nothing. So machines are able to now create new content and images and help to make decisions. For instance, if I were to ask someone in the audience to create an image of an artwork in the style of Louise Nevelson, a sculptor who was famous for creating art out of salvaged wood, that probably would be pretty difficult. But a properly informed and trained AI model can do this in under a minute, and that's what we call generative AI.Okay, so now moving this forward, in this advent of AI, we're really ushering in a new era. Welcome to the era of Copilots. Copilots are generative AI that assists humans with complex tasks. They're not used to replace humans, but really to be that copilot sitting side-by-side with us. And so they're augmenting human potential by unlocking productivity and unleashing creativity. They don't fly the plane on their own; they fly alongside us.For those of you on the line who might be familiar with Microsoft Office products, you can find Copilot which is embedded to assist every task within our Office products. So think about content creation and summarization — you can find that already as part of the Copilot experience. For the increasing technical users, you can start to create your own Copilot using Copilot Studio. So think about that as creating your own GenAI assistant to start lightening your load. But all of this is you're in control of the input as well as the output — and the value increases with shared usage and increasing complexity. After my section, I will be passing this over to my colleagues who will really cover that as AI and generative AI becomes a larger part of what we do and how we think about work, what are some of Microsoft's responsible AI principles that will secure in which how we're using AI and for our own productivity purposes.Okay, so throughout history, I think some of this can perhaps have all of us be on a bit of a pause. What does AI mean and how big of a transformation are we really looking at here? Throughout our history, the right tool in the right hands has always sparked transformation. From moving water across hundreds of miles via ancient aqueducts to putting the world's knowledge at our fingertips through smartphones, these tools have changed our lives. And now we're seeing the potential of the next great shift and that's the new era of AI. This shift isn't coming in a few years down the road, it’s not coming in a few months. In fact, you are in control. The opportunity is right now in terms of taking that AI into your own hands and making that work for you. So it is one of the most exciting shifts of our generation and we, along with our other colleagues at Microsoft, really see this impact taking shape for every individual, every process, every organization, and indeed, every industry.Okay, so now if you take a look at the slides and just imagine what could be the impact to your daily business or to your business units. What if software developers can code 55% faster? What if your employee attrition can be reduced by 20%? What if your marketing professionals can create content with much higher quality? And what if your customer service representatives can resolve 14 more customer issues per hour? And we're seeing those types of results through our customers with the usage of generative AI.And many of our customers have already created Copilots that meet the needs of their unique businesses. Let's take a snapshot of some of those customer-branded Copilots. Instacart and Carrefour, for example, have created new ways of shopping through conversational commerce. So for example, right now through Instacart, we as a consumer have different ways of discovering products. You can ask a question such as, "What kind of side dishes should I serve with lamb chops?" Now that's an entirely different way of engaging with a commerce channel than there was before.Coca-Cola has launched its own viral marketing campaign to endear the Gen Z crowds with limited-edition sodas. I know some of us in the audience perhaps are really responsive to that new innovation that really echoes with pop culture today. So imagine creating, testing, and getting feedback on a product in weeks or months versus in years — all on the basis of what GenAI believes consumer tastes will be like in the year 3000.Since March 2023, Air India's generative AI virtual agent called Maharaja has responded to 85% of 6,000-plus daily customer inquiries in four different languages without human intervention. Now this landmark has created a new understanding of AI-enabled conversation flow. In many ways, these patterns don't have to become intellectual property. It really is the implementation that makes it successful and makes it stick for an overall sustainable use of AI.Asahi Group has supercharged employee productivity and one of the outcomes is they are reinventing the approach to consumer segmentation, including reasoning over mass amounts of data to simulate consumer consumption behavior. Now, many of us probably have been in those studies to segment customers and those segmentations. It's difficult to change often, but now generative AI brings the ability to do that, to be able to reason over large amounts of data.And finally, Unilever uses AI-driven molecular chemistry simulations to test new compounds without manufacturing them. So it's in that in silico environment that now the material scientists with Unilever can use generative AI for natural language queries and automatically generate code for data scientist models, achieving quick unlock and real reliable scientific insights as well. So these customers have all taken generative AI and built it into their business processes. So our imagined future is becoming a reality.So as per the study and often with our customers, we often get the question of what is the ROI on generative AI investments? We encourage you to think about the value journey to begin with a solid base of individual usage and then moving that up. So at the base of that pyramid, it's about that solid foundation of individual usage. And then moving up the pyramid to optimizing departmental processes and impacting metrics that are important to your organizations. So let's take this pyramid from the bottom to the top.So the very bottom, it's individual value. Start by getting people to use the tools to unlock productivity from daily tasks like managing emails and collaboration, summarizing meeting outputs so that someone doesn't need to do that after the fact. This saves small-time segments that can be reinvested into the business.Moving on up to the middle of the pyramid is unlocking departmental value. As individuals learn, departments can discover broader use cases. For example, attorneys can use AI for contract drafting, HR for document-heavy processes, and sales and marketing for proposals and RFPs. This stage focuses on improving quality and efficiency within departments.Now we will move to the very top of that pyramid to unlock overall organizational value. This is the ultimate goal and it is to achieve business metrics such as cost savings, revenue growth, and operational efficiency. This involves leveraging AI for widespread use cases across multiple business units and within an organization to enhance customer interactions and reduce external costs. So for those of you familiar with Microsoft products, individual value unlocking could be found in the Copilot solutions that are already built into Office. For those departmental values, think about purpose-built Copilots, and some of you might be familiar with Copilot for Sales. So those are purpose-built and trained models for specific functions. And Copilot Studio is for a slightly more technical audience to use low-code and no-code options for citizen developers to start using data that they have access to in order to improve how they're functioning their days.So now I'd like to just bring your attention to a customer who has really achieved some of that organizational value. CarMax is a well-known reseller of used cars in North America. This example is not new, but it really resonates with me because I think about CarMax's challenge with having to create so many net new product detail descriptions. As all of those used cars are coming through, every car is a unicorn. And so they've realized that generative AI can really help to not only write the product descriptions but also start to summarize customer reviews into those product descriptions. That increases their SEO, as well as increases the discoverability of those products and helps consumers understand, “What are other people thinking and what kind of a car is it that I'm purchasing?”And so I think some of these solutions, some of the impacts, very much resonates with me as I think about 80% of the content goes through without a human being involved versus it used to take people lots and lots of time to create product detail pages that reflect on that very unicorn of a car. And it's estimated that would take an individual 11 years to do what Azure OpenAI has done in days. So it's not a new example. CarMax actually has been on this journey, which I also very much like to applaud CarMax for being in this journey already for a couple of years, of taking a piece of technology, scaling it, and now really seeing the material impact to revenue, cost, efficiency, and also consumer engagement.So as I conclude my section, perhaps many of you have been in the experimentation stage of generative AI. So you might be thinking, “How do we move from generative AI exploration to broader adoption?” And we recommend that you take a human-centered approach to prioritize the right teams and the right roles, and that requires an internal assessment of which teams and which roles are really ready for that. And so looking at that and then identifying the business processes within those teams to begin transforming.And then lastly, always keep an eye on the KPIs to ensure that the investment has produced positive returns, to be able to share the good news with others and help others scale this technology as well.Thank you very much for your attention. And now I'd like to transition over to my colleague Ranti, who will begin to talk about Microsoft's responsible AI principles.Ranti Okunoren: Thank you, Dina. Hello everyone. I'm Ranti and I'll be talking today about responsible AI and Microsoft's approach to responsible AI. Dina talked about shifts in technology and we're going through a shift right now with generative AI. And anytime there's a shift, there are always benefits and there are always risks. And when we launched our GPT-4 some of the risks and dangers we found with regard to GPT-4, we documented them in our transparency notes. Some of them are where you have errors and grounded outputs, drill breaks, which could be attacks from external sources, and sometimes in internal sources with harmful codes. We've also learned about people having worries and concerns about infringement claims, which my colleague will be talking about later — how Microsoft is mitigating that risk with our commitments and also especially within the consumer goods industry where we are dealing with a lot of customers and consumers. We are thinking about the risks of having generative AI or AI systems acting like humans and how to mitigate those risks.And so anytime there are benefits in technology, there are risks, and that’s one of the things I want to talk about today, about responsible AI and how responsible AI helps to mitigate those risks. If you think about what responsible AI really is, it’s the steps we take to ensure that our systems are trustworthy and uphold our principles. I'll be talking to you about Microsoft's six principles of responsible AI. Applying principles really is a matter of mindset and practice, so it has to be part of a culture and practice.We have six principles with regard to responsible AI, and I'll talk about them briefly. The first one is fairness. Fairness really is about mitigating bias. Fairness has to do with the allocation of resources and information in a way that is fair for all people. And so when we're developing systems, we think about ways to mitigate bias with the allocation of resources.The next is reliability and safety. When we're dealing with consumers in our consumer goods industry, we're thinking about safety, making sure that whatever systems we put together, whatever systems we're developing, designing, or deploying, we have to make sure those systems do not cause harm. Because like Dina mentioned, using the human-centered approach when we're developing the system, reliability is very important. And when we're thinking about reliability, we're thinking about how we're remediating any failures, remediating or mitigating any risks, and making sure that we have ongoing processes for monitoring privacy and security.In the generative AI era, data is a principal part of this whole shift. Making sure that we can keep data private and secure is going to be an important component.Inclusiveness is all about making sure that generative AI is accessible to everyone.Transparency and accountability are really very foundational. Transparency definitely plays a part in every state, whether we're designing, developing, deploying, making sure we're documenting, making sure we're communicating with stakeholders, and making sure that we're disclosing the risks, the limitations, and the capabilities of the systems we're developing. Accountability is also very, very fundamental in the sense that we have to make sure that, again, the human-centered approach, making sure the human is always in control and making sure that the human is basically at the center of everything.So if we put these principles in place, then we have a framework by which we work, and these are the things we want to consider when we think about the framework. The framework always starts with the principle. And then when we have this principle, you have the principles set down, and then you think about the corporate standards. And when we think about the standards, Microsoft released a second version of our standards, which is really how we think about putting it into practice because a principle is one thing, but actually what does it look like in practice? When you want to move from principles to practice, think about the how: how we make this work. I've talked to some of our customers and some of my colleagues with other companies, including retail entities and consumer entities, and they worry about basically starting from scratch. And I tell people, "Look, you don't have to start from scratch. Almost every organization has privacy policies already."So think about that. Think about the privacy policies you already have. Think about the security policies you already have. Think about the accessibility policies you already have. And if you're introducing generative AI systems, think about the gaps those policies have and just update them. And so if you already have your privacy policies, you just need to update them to accommodate for generative AI systems. And so that's where your goals, requirements, and practice come in terms of your policies, describing the how-to, step-by-step, and making sure you comply with the principles.Then after thinking about implementation, it’s really one step ahead when you're thinking about deploying the systems. And you put those processes in place and make sure people are trained. Remember I talked about transparency and how transparency is at every stage, whether you're designing, developing, or deploying. Training people and helping them to understand how the system works so they can use it correctly, helping them understand the limits of the system and also the benefits and the uses of the system. And training is something that I harp on a lot because I feel like all our employees in any organization are the last line of defense. If people know how to use the system, like Dina described, if people are trained and they use it every day, if they know how to use the systems, they understand the limitations and also the use cases, then they're better able to use it properly in a responsible manner.And so when we're thinking about implementation, we need to put processes in place to mitigate any risks. And then train everyone, train your employees, making sure that everyone understands the uses and the limitations of assistance we're putting in place and also tools to use those resources, making sure that everyone has access to those tools.Now, like I said, it's an iterative process where you're building, you're designing, you're developing. You also need to make sure that as you put the systems in place, people are trained. You need to make sure that you're constantly and continuously evaluating the system, making sure that what you set out to do in the beginning is what it's actually doing. And so monitoring is a big part of this whole framework and making sure that you're consistently monitoring, having a compliance team in place, monitoring and auditing as well, whether it's internal audits or external audits, making sure that we have this process in place. So if we think about this framework, the principles, the standards, the implementation and oversight, that is the entirety of the framework and that's Microsoft's approach to mitigating the risks of generative AI.And so when people ask about responsible AI and governance, we see this as a risk management framework. Basically, if you think about NIST that most of us are already familiar with, the framework is based on mapping, measuring, and managing — and that's the whole governance framework. And like I just discussed, in mapping, you identify the risks. When you measure, you quantify the risks, making sure you have the structures to mitigate those risks. Managing is when you continue to monitor and evaluate the risks. And so that's what the governance is all about. So identify rules and responsibilities of people that will be, remember it's a human-centered approach, making sure their roles, people who are at the center making decisions. Also documenting the processes, which is really the transparency principle. And a culture of risk management, which is putting processes in place, making sure your policies are up-to-date, and training your employees. And so that really is the entirety of the framework and ensuring your process, your system has your entity has the responsible AI program.This is Microsoft's approach and these are some of the things we want to consider when you're developing a responsible AI program in your entity, some things to consider with regard to governance. And I'll hand over to Nydia, my colleague, to talk about Microsoft's customer commitments.Nydia Galarza Rosario: Thank you, Ranti. Hi everyone. Today I'm excited to share with you Microsoft's AI customer commitments. As Ranti was mentioning, at Microsoft, we believe that every organization that creates or uses AI systems will need to develop and implement its own governance systems. That's why last year we announced three commitments, which are designed to support our customers on their responsible AI journey and ensure they can use our AI products with confidence. The AI customer commitments focus on three main areas, sharing our learnings about developing and deploying AI responsibly, creating an AI assurance program, and supporting our customers as they implement their own AI systems responsibly. Let's discuss each of them in more detail.First, sharing expertise. We're sharing what we're learning about developing and deploying AI responsibly and assisting our customers in learning how to do the same. We're committed to sharing this knowledge and expertise with customers by providing training curriculum and publishing the key documents we develop during our responsible AI process so our customers can learn from our experiences. These documents include our responsible AI standard, the AI impact assessment template, the AI impact assessment guide, and the transparency notes.Just to give you an example of the content of these documents, Microsoft transparency notes are intended to help customers understand how our AI technology works. The choices system owners can make that influence system performance and behavior and the importance of thinking about the whole system, including the technology, the people, and the environment. Customers can use our transparency notes when developing or deploying their own system or share them with the people who will use or be affected by the system.Our transparency resources also assist customers' understanding of how Copilot from Microsoft 360 and Azure OpenAI service process data and the extent to which additional information, if any, needs to be communicated to data subjects. Second, we have the Microsoft AI assurance program, which helps customers ensure that the AI applications they deploy on our platforms meet the legal and regulatory requirements for responsible AI. This program includes regulatory engagement support to help customers navigate regulatory issues based on our extensive experience in highly regulated industries. It also includes sharing our experiences implementing the AI risk framework published by the National Institute of Standards and Technology. The creation of a customer council to gather feedback to deliver relevant and compliant AI tools. And regulatory advocacy to actively engage with governments to promote effective AI regulation.And our third commitment is supporting our customers. And this means first that we're investing in dedicated resources and expertise in regions around the world to respond to customer questions about deploying and using AI responsibly. And we're also supporting our partners. Many of our partners have already created comprehensive practices to help customers evaluate, test, adopt, and commercialize AI solutions, including creating their own responsible AI systems. We're working with selected partners to leverage this expertise to assist our mutual customers in deploying their own responsible AI systems.And last but not least, the customer copyright commitment, by which we were the first AI provider to make clear that when a customer uses our generative AI services, we will defend most lawsuits brought by a third party claiming that the content the model generated for our customer infringes any IP rights.Let me talk a little bit more about the customer copyright commitment. As AI continues to transform the way we work and create, it also raises important questions about intellectual property. This innovation comes with the challenge of ensuring that the outputs generated by these tools do not infringe on existing copyrights. To address these concerns, Microsoft introduced the customer copyright commitment, which is designed to provide our customers with peace of mind. This commitment was implemented through an update to our commercial product terms last year.Our commercial licensing agreements include provisions that outline Microsoft's defense against third-party claims that our products infringe IP rights. By publishing the new commitment in the product terms, we expanded Microsoft's defense commitment to include output content of our paid commercial versions of Microsoft Copilot services and the Azure OpenAI service. No contractual change or action is needed on the part of our current commercial customers. However, as listed in this slide, our customers must have implemented the required guardrails and mitigations we have made available to be eligible for the benefits provided by the customer copyright commitment.And why this matters. This commitment reflects Microsoft's philosophy of standing behind our customers. For over two decades, we have defended our customers against patent claims, and now we're extending this protection to cover copyright claims related to AI-generated content. This is part of Microsoft's broader effort to balance the benefits of AI with the need to protect creative works, and it's a significant step towards ensuring that our customers can confidently use AI-powered tools without fear of IP infringement. By providing legal protection and promoting responsible AI use, we aim to foster innovation while safeguarding the rights of creators. Thank you for your attention. If you have any questions, we will be happy to answer them now.Dominguez: Thank you, Nydia, Dina, and Ranti, for that great presentation. We do have a few audience questions that we'll go through. This first one is for Dina, where should a CPG begin its journey of generative AI exploration, particularly when data quality is not in a good place?Zhou: Thank you, Liz, for that question. This is a common question from our customers. So while the ambition is there to explore the immense potential of generative AI, the journey can be a bit daunting, particularly when the data is not in a good place.So just a few recommendations in terms of how to start building that data foundation, and certainly this is not an exhaustive recommendation by any means. Start first by assessing your current state of data and think about how to improve the quality based on your current state of understanding. Conduct a thorough audit of your existing data sources. Identify critical data quality issues such as missing values, duplicates, and inconsistencies. Evaluate the reliability of your data collection process, those data sources, how you're collecting them, and what is the ongoing data cleansing process. Implement that data cleansing process. Develop automated data cleansing protocols, wherever that is possible, standardize data formats, ensure consistency across datasets, and utilize validation tools to detect and correct errors in real time. With the evolution of generative AI, we've also seen agents that have been developed in some of our latest announcements, and we start to have agents that can start to automate and help to be able to bring about what are common data challenges and help to automate the cleansing process of that as well.Then in terms of building out the rest of the data foundation, we'll go back to some of the points that were mentioned in the presentation. So as the current state assessment for data has been conducted, evaluate some of those low-hanging fruits for data quality management, and put in a data cleansing protocol that is repeatable. Then the next thing would be to prioritize use cases that have an immediate impact. It's pretty daunting to fix all of the data quality issues at once, so start to prioritize use cases that will make the biggest impact. Invest in data governance and management. Leverage external expertise if you find that that is necessary. And then lastly, foster a culture of data literacy so that you have democratized the ability to look at how data quality can be maintained going forward.This process of exploring generative AI is certainly very exciting, but all of the foundation is built on data and quality as that go-forward. So by improving data quality, prioritizing use cases, investing data governance, leveraging external expertise, and building this culture, you'll certainly start to see those benefits with generative AI, as mentioned.Thank you, Liz, and back to you.Dominguez: Great. And this one's for Ranti. What are some common mistakes that CPG organizations make when managing their use of AI and how can they avoid them?Okunoren: Thank you for that question. I think the first thing I will say is start with making sure that the whole responsible AI is developed at the leadership level. If the leaders adopt and agree to and are aligned with having a responsible AI culture and practice, it becomes a culture within the organization. Some of the mistakes that are made are that the leaders are not brought in, and all the stakeholders are not brought in. So making sure that all the important stakeholders are brought in, cross-functional groups, to make sure that there's fairness — having a wide array of options and contexts and points of views to make sure that you mitigate bias. The mistake people make is not having the right processes in place. We already have some policies. Most companies already have policies in place, but not going ahead to update those policies to reflect generative AI systems to address and mitigate the risks that come with generative AI. The other thing we've also seen is communication, which is part of the transparency principle, and also training. And I talk about training a lot. I talked about training earlier. Making sure the employees are properly trained because there's something about having systems in place and then having people actually use the systems that are deployed. When people are trained, they understand how to use the system and the risks that come with the system... people are able to use it responsibly and mitigate the risks that come with generative AI. The last thing I think I'll talk about is not really having a governance program process or framework in place. People talk about responsible AI principles, and we talk about governance. The principles are just a part of the governance framework. And so making sure you have your principles in place and then actually having the full framework that involves not just the principles, but the policies and the implementation processes, and also the monitoring. You can have a good system in place, but if it's not doing what it's set out to do, then it's not really accomplishing the goal. And it's never a one-and-done, it's a continuous iterative process.Dominguez: Great, thank you, Ranti. I actually have a follow-up for you on that AI governance topic. In terms of mitigating risks related to bias, fairness accountability and AI decision-making, what role does AI governance play particularly in areas like product recommendations or pricing strategies?Okunoren: Yeah, definitely keeping with the principles. When you're thinking about pricing, merchandising and you think about fairness, for example, work with cross-functional teams. And to mitigate bias, you have to have a diverse set of people working on the products, working on the pricing, having different perspectives. And so cross-functional teams would be the way to mitigate that bias. Cross-functional teams in the organization. Like I tell people, when you have just IT people thinking about AI systems that really won't work. You need to have your attorneys in the room, you need to have your compliance people in the room, you need to have your product people in the room, you need to have your IT people in the room. And so you have a diverse perspective of people. And that way, those systems, you're able to mitigate the risks that come with AI.Dominguez: Great. Thank you so much. And Nydia, I have a question for you here. What strategies can CPGs use to build internal processes and frameworks that ensure AI solutions are implemented effectively and ethically, especially when scaling across global markets?Galarza Rosario: Thank you, Liz. This is a great question. And talking about governance. It's all about governance. So the optimal strategy involves establishing a comprehensive governance framework that includes clear roles and responsibilities, regular training on responsible AI practices, and continuous monitoring of AI systems. Additionally, leveraging resources from other companies can be very beneficial. I encourage you to visit Microsoft's responsible AI website. This site provides tools, practices, templates, and a lot of information to help establish good governance foundations for AI compliance.You also have to think that any framework should be adaptable to different markets and regulatory environments to ensure that they scale globally. So it is also very important to stay informed about regulatory developments that may impact your industry. So definitely engage with your counsel to get more information about any possible regulations and include that into your governance framework. Thank you.Dominguez: Thank you. Okay, we have another question here. Dina, this one's for you. Any favorite AI-first companies or brands in CPG?Zhou: Thank you for that question. It's a hard one to answer because, from our perspective, there's not a single company that is not actively exploring AI and generative AI in that respect. So it's a difficult one to answer in terms of is there a favorite one. Instead, if I could just pivot the question a bit in terms of are there some favorite use cases or favorite functions in which we have seen companies actively exploring? And it really starts in the sales and marketing space. And I think that also was reemphasized by the CGT report because it's consumer-facing, because there's some active content creation and summarization, which generative AI was really first created to do and it was doing really well. So I would say those functions are the functions that are charging ahead and very, very experiential in terms of driving content generation and new ways of engaging with the consumer, such as the Carrefour example, such as Instacart.Another area that I've seen just a ton of ambition around is of course R&D because there's just a massive amount of data, a massive amount of curation discovery that GenAI has the ability to help. And then the third area that I have seen is in supply chain manufacturing. Again, now with just so much disruption that is happening and sometimes it's multilayered level of disruption, generative AI can help companies to think ahead about what could be the mitigation strategy that could get ahead of a disruption that is coming up next.So thank you for that question and I hope I'm answering a bit that still resonates, but know just that there's not a single customer that is not exploring and everyone is doing an amazing job in terms of thinking about how that could apply to their business, but making it in a responsible way with clear governance is our recommendation to you. Dominguez: Great. Thank you so much, Dina. Well, I'm afraid that's all the time that we have for today, so I do want to go ahead and wrap things up. If we didn't get to your question today, we'll be sure to have someone follow up with you. I'd like to thank our speakers for giving us their time and their subject matter expertise. As a reminder, you can read the full sales and marketing report at ConsumerGoods.com. I'd like to thank Microsoft for sponsoring today's webinar. And once again, the webinar recording will be available on demand shortly on ConsumerGoods.com. Finally, thank you to our attendees for giving us your time today. We do hope that you found it worthwhile. Have a great rest of your day.

Related Topics