Google Cloud Next 2024 Michelob Ultra Arena Opening Keynote: The New Way to Cloud Philip. >> You've got people working anywhere, with data, apps, and devices everywhere, and hackers who can't wait to attack. It's time to implement a network security platform built for zero trust, powered by AI. That's how you'll have the intelligence you need to prevent the latest threats, and that's how you'll secure whatever, whenever, wherever, with less complexity. Move your business forward and never look back. Google Cloud Next 2024 Michelob Ultra Arena Opening Keynote: The New Way to Cloud >> Where could reinvention take your business? Accenture, let there be change. ♪ ♪ ♪ >> At Google, AI isn't a tomorrow thing, it's a today thing. And sure, we can ask for answers to life's tough questions, but today, we can also ask it to do tough things, like ship that wood more efficiently, use satellites to reduce methane emissions, turn DNA into code to make drought-resistant corn? Oh, cool! It can spot and fill potholes, spot disease earlier, spot asteroids to protect Earth. It can create apps with just words in just hours. Today, Google AI can scan 100,000 lines of code in two minutes to spot and fix bugs. It can translate from code to code and is one step closer to speaking 1,000 languages, even whale. Today, AI impacts lives for the better and understands the world the way you do. Our AI spots threats seven times faster, Understands speech sentiments, helps prevent vision loss, predicts weather ten days out in two minutes, translates thousands of pages in seconds, detects lead pipes to keep drinking water clean, eliminates paperwork so care is more human, generates from text, visuals, audio, and video, creates entirely worlds from your imagination, creates this thing and that thing, a new thing, the new way to cloud. >> ANNOUNCER: Please welcome our CEO of Google Cloud, Thomas Kurian. ♪ (Applause) >> THOMAS KURIAN: Wow! Welcome, everyone, to Google Cloud Next. It's been less than eight months since Next 2023, but we've made A world of progress with all of you. We've introduced over 1,000 product advances across Google Cloud and Google Workspace, expanding our infrastructure footprint to 40 regions, including Dammam and Johannesburg, and announcing six new subsea cable investments; being recognized as a leader in 20 top industry analyst evaluations in just the last year, and more than a million developers now use our state-of-the-art Generative AI in popular tools including Google AI Studio, Vertex AI, Colab, VS Code, JetBrains, Replit and many more. Our Generative AI trainings have been taken millions of times, and our services partners have taken over a half million gen AI courses. And we've seen great success with customers and partners, making us the fastest growing cloud provider. Let's hear Sundar Pichai put this interesting time in our industry into context. ♪ >> SUNDAR PICHAI: Hi, everyone. Welcome to Las Vegas and to Google Cloud Next. I'm told I'm on a very large screen right now, so I'm just going to imagine this is my audition to play the Sphere. We were last together in San Francisco in August. As we all know, months can feel more like decades in cloud time. Just think about it in the context of generative AI. Last summer, you were just beginning to imagine how this technology could transform businesses; and today, that transformation is well under way. We have seen incredible momentum in the cloud business since then. Last quarter, Cloud was at the $36 billion annual revenue run rate. That's five times our run rate five years ago. Congrats to Thomas and team, and a huge thanks to all of our customers and partners. I want to highlight just a few reasons Google Cloud is showing so much progress. One is our deep investments in AI. We have known for a while that AI will transform every industry and company, including our own. That's why we have been building AI infrastructure for over a decade, including TPUs, now in their fifth generation. These advancements have helped customers train and serve cutting-edge language models. These investments put us at the forefront of the AI platform shift, and we are proud that today, more than 60% of funded Generative AI start-ups or nearly 90% of gen AI unicorns are Google Cloud customers. We also continue to build capable AI models to make products like Search, Maps, and Android radically more helpful. In December, we took our next big step with Gemini, our largest and most capable model yet. We have been bringing it to our products and to enterprises and developers through our APIs. We have already introduced our next-generation Gemini 1.5 Pro. It's been in private preview in Vertex AI. 1.5 Pro shows dramatically enhanced performance and includes a breakthrough in long-context understanding. That means it can run 1 million tokens of information consistently, opening up new possibilities for enterprises to create, discover, and build using AI. There is also Gemini's multimodal capabilities which can process audio, video, text, code, and more. With these two advances, enterprises can do things today that just were impossible with AI before. For example, a gaming company could now offer video analysis to support players, including tips to improve performance; or an insurance company could combine video, images, and text to create an incident report and automate the claims process. We are excited to bring our decades of innovation and research to our customers as you pursue your own opportunities with AI to grow and evolve. You will hear from hundreds of companies using Google AI to do just that here at Cloud Next. That includes four CEOs of iconic companies from around the World: David Solomon, who will share how Goldman Sachs is using AI to synthesize complex information quickly, without compromising quality; Ola Kallenius, sharing Mercedes-Benz building new experiences for drivers. And Dara Khosrowshahi will explain how Uber is using AI to empower employees and improve experiences, and Nikesh Aurora is useing Generative AI o help find the security needle in the data haystack. We have had the honor of working with these companies for years and we are looking forward to this next chapter and to seeing how AI will help all of you transform your businesses and innovate today and for the future. With that, I'll turn it back to Thomas. Enjoy the rest of the show. (Applause) >> THOMAS KURIAN: Thank you, Sundar. Since last year's Next, we've announced partnerships with hundreds of leading customers. Companies like Anthropic, AI21 Labs, Contextual AI, Essential AI, and Mistral AI are serving or training their own models on Google Cloud. Leading enterprises like McDonald's, Deutsche Bank, the Mayo Clinic, HCA Healthcare, U.S. Steel and many, many more are building new gen AI apps on Google Cloud. Yesterday, Cintas shared how they used Gemini models in Vertex AI to search their extensive library of contracts, documents, and products. Looking at Next by the numbers, There's 30,000 people live here to today, 2,000 Google experts, over 2,500 partners, and nearly 400 sponsors. A special thanks to our luminary spon sponsors: Accenture, Capgemini, Cognizant, Deloitte, and Palo Alto Networks. (Applause) and partners sharing their gen AI success stories at this e event. Today, we're going to focus on how Google is helping leading companies transform their operations and become digital and AI leaders. We call this the New Way to Cloud. We have many amazing advances, starting with infrastructure. We are the first leading hyperscaler to market with Intel's fifth-generation Xeon Processors, continuing our leadership in price performance. Google Distributed Cloud and Edge advances bring Computer, Data, AI, and ISVs to highly confidential and edge workloads. Our Cross-Cloud Networking allows you to connect any service, including our AI service, to any application on any cloud, securely, with lower cost and latency, including all of our AI Services; databases have enhancements to AlloyDB, Bigtable and Firebase to improve relia reliability, performance, and scalability. BigQuery is evolving to be a unified end-to-end platform for data and analytics with a unified user interface, unified security, and governance. Security products include the introduction of Chrome Enterprise Premium browser for greater security and control for enterprise, along with major advances in our next-generation firewall, Cloud Armor, and confidential compute. Google Workspace is introducing enhanced security, extension models, and a new Chat plus Meet offering with the best voice quality in the industry in Google Meet. And that's just the tip of the iceberg. There's a whole lot more! Our biggest announcements are focused on Generative AI. Customers have quickly gone from AI helping to answer questions to AI making predictions, to Now building Generative AI agents. Agents are intelligent entities that take actions to help you achieve specific goals; for example, helping the shopper find the perfect dress, helping an employee pick the right health benefits, or having nursing staff being helped to expedite patient handoffs when shifts change. Agents process multimodal information simultaneously, can converse, reason, learn, and make decisions. Agents can connect with other agents, and with humans, and they will transform how people interact with computing devices and the Web itself. Our customers are building early versions of AI agents using our AI infrastructure, our Gemini Models, Vertex AI, Workspace, and Google Cloud. And AI is transforming many industries. Goldman Sachs is at the cutting edge of technology in financial services. Please welcome their Chairman and CEO, David Solomon. ♪ >> DAVID SOLOMON: It's been a fascinating journey with AI technology. We've seen rapid Innovation, and we're now at a stage where practical application of the technology is showing the potential to yield real value for the business. At Goldman Sachs specifically, our work with Generative AI has really been focused on three key pillars: Enabling business growth and enhancing client experience; improving developer productivity; and increasing operating efficiency across the firm. We want to be thoughtful in how we approach and harness this technology, both as a mechanism to drive efficiency and as an enabler of disruption and differentiation for our clients. We're already seeing signs of promise in a few areas of our experimentation, and we're very optimistic about that. There's evidence that Generative AI tools for assisted coding can boost developer efficiency and productivity by as much as 40%. And we're exploring different ways to use AI, whether it's to summarize public filings, extract sentiment and signals from corporate statements, or to gather and interpret information, like earnings reports. We have a world-class team of engineers working on this, and we've already learned a lot. I'm super excited to see what's next. (Applause) >> THOMAS KURIAN: Thank you, David. We're grateful for your partnership and progress building with our complete, open, and integrated AI platform, which includes our A iI hyper computer, an integrated and optimized system that leads the industry in cost, performance, productivity, and scale for AI training and serve i ing. Our foundation model, Gemini, which has advanced reasoning skills with multimodal infor information. We are the only cloud to offer widely used first-party, third-party, and open-source models. Vertex AI to access, tune, augment, and deploy custom models and to build agents. Gemini for Google Cloud, which helps with every aspect of Google Cloud, including writing code for any environment, and Gemini for Work Workspace, which is the agent we built right into Gmail, Google Docs, Sheets, and more, with enterprise-grade security and privacy. And we offer integrated partner solutions at every single layer of the stack. Let's start by looking at our AI infrastructure. Please welcome Amin Vahdat. ♪ >> AMIN VAHDAT: Supporting the promise of Generative AI, from building world-leading multimodal models to serving them requires harnessing computing and data at a scale never imagined before. In fact, we are seeing computing demands of large language models grow by up to a factor of 10 times a year. That is why we designed and built AI Hypercomputer, an orchestra of hardware and software, from programming languages, to compilers, runtime, serving stacks to the chips and networks that make it all possible at an unprecedented scale. This system-level integration is up to 2x more efficient at scale relative to baseline solutions that simply deliver raw hardware and chips. The combination of our vision and execution has been broadly recognized. As one example, we were just recognized as the leaders in Forrester's AI Infrastructure Solutions wave, in offering and in strategy. But we're continuing to introduce new enhancements at every layer of the stack. First, we have the leading portfolio of performance-optimized accelerator, including Google-designed TPUs and NVIDIA GPUs. Today, we're announcing several enhancements to our GPU portfolio, including the upcoming general availability A3 mega, Powered by NVIDIA H100 tensor core GPUs with twice the network bandwidth per GPU compared to A3 instances. The rate of progress with NVIDIA GPUs is truly astonishing. Today, we are announcing support for NVIDIA's newest Grace Blackwell generation of GPUs, coming to Google Cloud very early in 2025. The NVIDIA B200 and GB200 chips are both powered by next-generation NVIDIA networking. GB200 will be among the most advanced chips on the planet and will require liquid cooling to operate at peak efficiency. Finally, we're also announcing the general availability of TPU v5p, our most powerful and scalable TPU. Our latest generation TPU pod consists of 8,960 chips, interconnected to support the largest scale ML training and serving. Our TPU v5p pods have 4x the compute capacity per pod, compared to our previous generation. And with TPU multislice technology, you can simultaneously and transparently run across multiple pods for your most demanding workloads. In fact, we train and serve our latest Gemini models on our TPUs. To improve A I training and inference speeds, we've added a number of enhancements to our storage products. Specifically, today we're accelerating inference with Hyperdisk ML, our n next-generation block storage service optimized for AI inference and serving workloads. It accelerates model load times up to 11.9x compared to common alternatives and offers over 100 times greater throughput per volume versus compet competitors. And for ML developers, we continue to optimize open and leading ML frameworks, like Jax, PyTorch and TensorFlow for use with both Vertex and GKE. Jax, an open-source partnership that includes NVIDIA, AMD, Intel, and others, has emerged as the leading framework for the most advanced ML practitioners. It allows developers to focus on the training logic declaratively, rather than partitioning code manually. Our XLA compiler then generates code for a range of GPU and TPU targets. In fact, over the last year, the use of GPUs and TPUs on GKE has grown more than 900%. Finally, with data-intensive workloads like AI, efficient resource management is vital. Today, we are launching two new options for Dynamic Workload scheduler, our resource management and job scheduling t tool. Calendar mode for start time assurance and flex start for optimized economics. In combination, these make a massive impact on system throughput and on cost efficiency. Overall, we find that our A I Hypercomputer can run with more than twice the effective efficiency relative to baseline hardware-only techniques. Leading companies like Anthropic have trained and served models on our AI Hypercomputer, and Kakaobrain, part of Korean technology company Kakao Group, has built a large-scale AI language model that is the largest Korean-language-specific LLM in the market with 66 billion parameters. They have also developed a text-to-image generator called Karlo. We are also bringing AI closer to where the data is being generated and consumed, to the Edge, to air-gapped environments, to Google Sovereign Clouds and Cr Cross-Cloud, through Google Distributed Cloud. Today, we are announcing that we are adding NVIDIA GPU support, adding GKE capabilities for AI workloads and enabling a variety of OpenAI models to run on GDC, including Gemma and Llama. We are bringing the power of Vector Search to allow search and information retrieval applications for private and sensitive data. For your most stringent regulatory requirements, we deliver fully air-gapped configurations with local operators of your choice. You have complete control of your data, including location, encryption, and access control. And GDC now has both Secret and Top Secret accreditations. Mobile provider Orange operates in 26 different countries where local data must be kept in each country. They are using AI on GDC to improve network performance and to deliver super-responsive translation capabilities. AI Systems run in concert with traditional computing sys systems. Google has a long history of designing processor systems, including TPUs for AI since 2013; VCUs for efficient video transcoding, and Tensor processors in phones. And today, we're thrilled to share something very special with you, but there's another person who's just as excited as I am. ♪ >> THOMAS KURIAN: We're delighted to announce the Google Axion Processor. It's our first custom ARM-based CPU designed for the datacenter, and it will be available in preview later this year. Google Axion -- (Cheers and applause) Google Axion combines Google's expertise with ARM's latest compute core designs to better deliver up to 50% better performance and up to 60% better energy efficiency than comparable current-generation X86-based VMs. That's why we've already started deploying at-scale Google services on ARM-based instances, including Spanner, BigQuery, GKE, Google Earth Engine, and the YouTube Ads platform. Now, let's dive into Vertex AI, our fast-growing enterprise AI platform. In our Vertex AI model garden, you can access over 130 models, including the latest versions of Gemini, partner models like Claude from Anthropic, and popular open models, including Llama, Gemma, and Mistral, in a variety of configurations. You choose the best model for your use case, budget, and performance needs, and switch between models as you need. Today, we're taking Gemini 1.5 Pro into public preview. (Cheers and applause) Gemini offers the world's largest context window, with support for up to 1 million tokens. With Gemini 1.5 Pro, customers can now process vast amounts of information in a single stream, including one hour of video, 11 hours of audio, codebases with well over 30,000 lines of code and over 700,000 words. We are enhancing Gemini 1.5 Pro with the ability to process audio, enabling cross-modality analysis. For instance, you can use it to search in audio and video content; for example, find a timestamp in a baseball game video, where a commentator says "It's outta here!". We've seen some amazing examples of what people can do with this large-context window. Sundar mentioned a few, and others include: A university professor is using it to extract data from a 3,000-page document with text, data tables, and charts, in just a single shot; a financial services company is using it to search, analyze, understand, and answer questions on complex merger-and-acquisition documents; a start-up founder is using it to understand code bases, to identify critical issues, and implement fixes; and a sports technology company that provides AI-enabled cameras lets users query their videos to evaluate plays or find key moments. Keep them coming! Today, Claude 3 Sonnet, Claude 3 Haiku, Anthropic's state-of-the-art models, are generally available on Vertex AI, with Claude Opus available in the coming weeks. We're also announcing the availability of CodeGemma, a fine-tuned, lightweight, open model designed for coding from the same technology used to create Gemini. With these additions, Google Cloud continues to be the only cloud provider to offer widely used first-party, third-party, and open-source models. Vertex AI can be used to tune, augment, manage, and monitor these models. We provide a variety of tuning and customization techniques: Fine tuning, reinforcement learning with human feedback, low-rank adaptation, distillation and mo more. To make it easier for all of you, today we are introducing supervised, adapter-based fine tuning, so you can customize Gemini in an efficient, lower-cost way. To augment models, Vertex AI provides managed tooling to connect your model to enterprise applications and database using extensions and function calling. Vertex also provides Retrieval Augmented Generation, combining the strengths of retrieval and generative models to provide high- high-quality, personalized answers and recommendations. Vertex can augment models with up-to-date knowledge from the Web and from your organ organization, combining Generative AI with your Enterprise Truth. Today, we have a really important announcement. You can now ground with Google Search, perhaps the world's -- (Applause) perhaps the world's most trusted source of factual information with a deep understanding of the world's knowledge. Grounding Gemini's responses with Google Search improves response quality and significantly reduces halluc hallucination. Second, we're also making it easy to ground your models with data from your enterprise databases and applications and any database anywhere. Once you have chosen the right model, tuned it, and connected it with your Enterprise Truth, Vertex's MLOps can help you manage and monitor models. For instance, we're introducing new prompt management tools today to help developers collaborate on prompts with notes and status, to track changes, compare response quality from different prompts. You can evaluate and improve response quality with the general availability of our automatic side-by-side feature, which provides explanations of why one response outperforms ano another, and certainty scores for model accuracy. Our Rapid Evaluation feature, which is now in preview, lets you quickly evaluate models on smaller data sets. An example, Generali Italia, Italy's largest insurance provider, used Vertex AI to build a model evaluation pipeline that helps its ML teams quickly evaluate performance and deploy models. To make it easy for all of your developers to access Gemini and other models on Vertex AI, we've built integration with many partners, including Hugging Face and Replit. They now allow you to access the models within ISV appli applications, in popular development tools, and from applications running on any clo cloud. With all this, Vertex AI is the only AI platform to provide a single platform for model tooling and infrastructure. Now let's look at the types of agents customers are building on Google Cloud use iing Generae AI. First, customer agents. You know, similar to great sales and service people, customer agents are able to listen carefully, understand your needs, recommend the right products and services. They work seamlessly across all your channels -- the Web, your mobile app, your point of sale, and your call center, and they can be integrated into products experiences with voice and video. Mercedes-Benz is working with us on customer agents to help people in their amazing cars. Let's hear from their CEO, O O Ola Kallenius. ♪ >> OLA KALLENIUS: At Mercedes-Benz, we want to offer our customers an exceptional digital experience. That's why we're equipping our cars with high-end computers. Each car should only get better over time, just like a good wine. And with the power of Google Cloud and AI, we will make the user experience even more personalized. Our partnership across Google helps us build more intuitive and customized experiences. Last year, we announced our partnership with Google Maps, and today, more than 3 million customers are using Google Places in their Mercedes cars, and we are ap applying Google Cloud AI across a number of other use cases, ranging from a smart sales assistant, improving customer service in our call centers, and optimizing our marketing. The sales assistant, for example, helps customers to seamlessly interact with Mercedes, when booking a test drive or navigating through Mercedes' offerings to find their next favorite vehicle. And now, we're exploring further opportunities to work with Google Cloud AI, such as next-level navigation features. In addition, we are partnering on one of the most exciting technology topics in our industry -- automated driving. This beautiful car right here is equipped with a level 3 system for conditionally automated driving. We were the first manufacturer to get it certified in Germany, California, and Nevada. For our Next-generation internal development and test platform, we will use Google Cloud as the backbone, helping us to become even more efficient and flexible in our product development. And Google Cloud's expert knowledge in processing massive amounts of data and scaling AI workloads will ensure that our cars get even more intelligent and AI driven. Partnering with the very best in their respective fields is an important part of our software strategy, and Google is the perfect example of that. With Google Cloud, Mercedes-Benz is building new ways to deliver the most intelligent vehicles to our customers and to create personalized, intuitive experiences. We're really excited about working together. Thank you for having me. >> THOMAS KURIAN: Thanks, Ola and team. In addition to these great innovations, we are very excited to collaborate with Mercedes-Benz and work towards bringing Gemini models to their in-vehicle conversational system. Another great example is Samsung, who is deploying Gemini Pro and Imagen on Vertex AI to their Galaxy S24 smartphones. Users can now take advantage of amazing features -- text summarization, organization, and magical image editing. Gemini Pro also provides Samsung with critical Google Cloud features including security, safety, privacy, and data compliance. We're inspired by the agents our customers are creating using our Generative AI platform and our models. Intercontinental Hotels Group will launch a travel planning capability to help %ach of you, their guests, plan their next vacation. ADT is building an agent to help customers select and set up home security systems. Verizon gives agents better recommendations. Magalu, one of Brazil's largest retailers, has put Generative AI right at the heart of its customer service. ING built a chatbot to enhance self-service and improve answer quality. Target uses AI on the Target app and site to personalize offers for Target Circle and Starbucks at Drive Up. Minnesota's Department of Public Safety helps non-English speakers get licenses and other services with realtime translation. Best Buy is building an assistant that will help troubleshoot product i issues, reschedule or combine order deliveries, or manage soft software. Discover Financial Services is using search and synthesis across detailed policies and procedures during customer service calls. And Orange's French-language agent is grounded in support knowledge, transforming their Help and Contact site and customer experience. OPPO and OnePlus, leaders in smart devices, are incorporating our Gemini models and Google Cloud AI into their phone to deliver innovative customer experiences, including news, audio recording summaries, AI toolbox, and much, much more. You know, the opportunity for customer agents is tremendous. To help each of you build customer agents faster, we're introducing Vertex AI Agent Builder. You can now create customer agents that are amazingly powerful in just three key steps. First, you can use Gemini Pro to create free-flowing, human-like conversations with text, voice, images, and video as inputs and personalize them with custom voice models. Second, you can use natural-language instructions to control the conversation flow and guide it on specific topics you don't want it to discuss, such as current events, in the same way that you train your human agents. You can also control when it hands over to a human agent with transcription and summarization of its conversation history to make these transitions extremely smooth. Third, you can improve response quality with Vector-based and keyword-based search to connect your internal information and the entire Web. You can also use Extensions to complete tasks for customers, like updating contact information, booking a flight, ordering food, and many mo more. And you can integrate enterprise data from operational databases like AlloyDB, predictive analytics with BigQuery, and SaaS Applications like ServiceNow. Let's take a look at an example of a customer agent in action. Please welcome Developer Advocate Amanda Lewis. (Applause) >> AMANDA LEWIS: Thank you, Thomas. So, last night, I was watching a video of this band, and I loved the keyboard player's shirt. So, I was thinking, "I'd really like to be wearing that shirt tomorrow night," but can I find it in my size and in time to be rockin' it at the concert here in Vegas? Let's head over to my favorite store, Cymbal Fashion. They just launched a customer agent, and it leverages Gemini and Vector Search to deliver a seamless shopping expe experience. "What can we help you find?" Well, I'd like that shirt, but I guess I have a few other specifications as well, so, "Find me a checkered shirt like the keyboard player is wea wearing. I'd like to see prices, where to buy it, and how soon can I be wearing it?" I'm going to include the video. Now, the customer agent is using Gemini's multimodal reasoning to analyze the text and video to identify exactly what I'm looking for. Then, Gemini turns it into a searchable format. Okay. How cool is this? It found the checkered shirt I'm looking for! Right? And some other great options, in no time. And that's because these results harness Google's trust ed search technologies, which ensures customers, like me, get the right results in record time. The suggested products are grounded in Cymbal Fashion's inventory and historical performance data to make sure customers leave happy and with that purchase in hand. Okay, the first one is perfect. It looks like I can have it delivered in four days or pick it up nearby today. Like I said, I want to be wearing it tomorrow night, so I'm going to go with the local store, to be safe. Of course. It never fails. They only have three left in my size. I don't want to miss out on wearing this shirt, so I'm going to go ahead, give the store a call, and ask them to set it aside for me. But first, let me tell you what's happening behind the scenes. Cymbal Fashion's customer agent is using Google Cloud's full suite of AI capabilities to offer customized support interactions. Facilitate transactions like purchases and returns, and ensure that I'm receiving the most up-to-date information in realtime. I'm so close to having this shirt for the concert. Let's give the store a call. (Ringing) >> CUSTOMER AGENT: Hi, there. This is the Cymbal Fashion Customer Agent at South Las Vegas Boulevard. Am I speaking with Amanda? >> AMANDA LEWIS: Yes! This is Amanda. >> CUSTOMER AGENT: Great. Thanks for reaching out, Amanda. I see you had a session on another device. I've sent you an SMS message with a link to our live chat companion. If you would like to switch to chat, please click the link. How can I help you today? >> AMANDA LEWIS: I'd like to purchase the shirt in my cart with the card I have on file. >> CUSTOMER AGENT: Absolutely. I see you're also a Cymbal Fashion Rewards member. Looks like you have a 20% off voucher available to use. Would you like to apply it to this purchase? >> AMANDA LEWIS: Yes, please. That would be great. >> CUSTOMER AGENT: The shirt you're purchasing goes well with these items, also available for pickup in your preferred size. Would any of these be interesting to you? >> AMANDA LEWIS: Absolutely. Please add the white shirt and the boots to my cart. >> CUSTOMER AGENT: Great. Your total is $203.73. Okay to proceed with the card on file? >> AMANDA LEWIS: Yes. >> CUSTOMER AGENT: Your purchase is confirmed. Do you need anything else today? >> AMANDA LEWIS: No. I'm all set. Thank you! Incredible. >> CUSTOMER AGENT: Thank you for shopping with Cymbal Fashion. You'll get a text when the items are ready for you. (Applause) >> AMANDA LEWIS: In less than five minutes, I was able to find and order the shirt I wanted and an outfit to match. I'm ready for the concert. Back to you, Thomas. (Applause) >> THOMAS KURIAN: Thank you, Amanda. That was awesome. Now let's talk about employee agents. Many of our customers are building employee agents using Gemini for Google Workspace and Vertex AI to help their employees be more productive and work better together. For hundreds of millions of people around the world, Uber is an essential mobile app that makes it easy to get a ride or a home delivery. There are many complex processes that must run smoothly behind the scenes for Uber to keep providing its magical user experiences. We talked with Dara Khosrowshahi, the CEO of Uber, and learned more about how they're building AI agents to help everyone. ♪ >> DARA KHOSROWSHAHI: Generative AI opens up opportunities across Our business, but the opportunity I'm particularly excited about is developer and employee productivity. Our partnership with Google Cloud is helping us create a stronger and more empowered workforce. Gemini for Google Workspace helps us save time on repetitive tasks, frees up developers for higher-value work, reduces our agency spending, and enhances employee retention. We're also use ing Google Cloud's AI platform to build our own AI agents to help our support teams better help our customers. For frontline support, we've launched new tools that intelligently summarize user communications and surface contexts from previous interactions. Anything that helps our teams go faster, helps our platform go faster, so our customers can get where they want to go or what they want to get delivered, all at Uber speed. With Google Cloud, Uber's creating new ways to empower our workforce and better serve our customers. Thanks for having me. Vice President and General Manager of Google Workspace, Aparna Pappu. >> APARNA PAPPU: Hello, everyone! Gemini for Google Workspace is our AI-powered agent built right into Gmail, Docs, Sheets, and more. Gemini doesn't just help our customers be more productive, it's a member of their team. Need a quick answer on something? Just ask Gemini. Need a world-class notetaker in an important meeting? Bring Gemini along. Need insights from a dozen analyst reports? Gemini is your research partner. Customers really appreciate the quality of output that Gemini for Workspace prov provides. More than 70% of Help Me Write users in Google Docs and Gmail accept Gemini's suggestions; and more than 75% of people who have created images and slides insert them into their presentations. We're seeing the impact of this every day with our customers. Pepperdine University has students and faculty who speak many languages. Gemini and Google Meet enables realtime translated captioning and notes. Pennymac is accelerating recruiting, hiring, and new employee onboarding. And Woolworths, the leading retailer in Australia, is seeking to unlock greater productivity for over 10,000 administrative employees by providing them with Gemini for Google Workspace. Now, today, we have quite a few announcements for Gemini for Workspace. First, many customers have asked for a way to get AI-powered Google Meet and Chat company-wide at a lower price, particularly to move away from Zoom. In a recent benchmarking study that tested the top video conferencing applications, Google Meet outperformed Microsoft Teams, Zoom, and WebEx in overall video and audio perfo performance. (Cheers and applause) So, we're announcing the AI Meetings and Messaging Add-on with Take Notes for Me, chat summarization and realtime translation, and it will only cost $10 per user per month. Second, we're making AI-powered data protection available with our brand-new AI Security Add-on. Now, Workspace admins can automatically classify and protect sensitive files and data using privacy-preserving AI models trained for their organization. Third, Gemini and Chat can now catch you up on long conversation threads, summarize decisions, and be a realtime creative partner. Let's have Product Lead for Workspace Collaboration Apps Kristina Behr give you a quick look at Gemini for Workspace. >> KRISTINA BEHR: Thanks, Aparna. There are lots of ways Gemini for Workspace is helping our customers save time. Let me show you the magic of one example in action. I've been asked to evaluate two proposals for a new payroll system. I open up my drive and I see that there are two statements of work, or SOWs, that outline the vendor proposals. Google Drive is AI ready without any additional prework, and everything here is already protected with our industry-leading security. All I need to do to get started is drag and drop these two files over to the Gemini side panel, which is currently in preview. In just a few seconds, Gemini for Workspace will give me an outline of these proposals so I can have a good idea of what they're about, and I could go even further and as ask: "Compare the price of the two offers." Each of these documents is over 70 pages. It would have taken me hours to review these docs, but instead, Gemini is going to help me find a clean answer to save me a ton of time. And as you can see here, really quickly, we got the price of the two offers. One of the things you'll notice about Gemini in Docs is it will proactively give me a summary of this file, so I can get a sense of what's going on without needing to take the time to read the whole thing. But before I proceed with this vendor, I need to be sure that no compliance issues exist. And I'm going to be honest, I have not memorized every rule in our Compliance Rulebook because it is over 100 pages. I would have to need to scour the 80 pages of this proposal and compare it manually with 100 pages of the Rulebook. So, instead, in the side panel, I ask, "Does this offer comply with the following?" And I'm going to "at" mention our Compliance Rulebook, hit Enter, and see what Gemini has to say. Okay, so, interesting. Gemini has found an issue because the supplier does not list their security certifications. Because Gemini is grounded in my company's data with source citations to specific files, I can trust this response and start to troubleshoot before selecting a vendor. Gemini for Workspace was terrific in this example, saving me a lot of time without compromising on accuracy. And this technology can be applied in so many other use cases: A sales team analyzing RFPs; a recruiter developing interview questions; and so much more. Back to you, Aparna. >> APARNA PAPPU: Thank you, Kristina! (Applause) Many of our customers are building employee agents to automate workflows that are tedious and repetitive. For example, Bristol Myers Squibb is transforming its document processing for clinical tr trials. Something that took scientists weeks now takes them just minutes to get a first draft. Dasa, the largest medical diagnostic company in Brazil, is helping physicians detect relevant findings in test resu results. Sutherland, a digital transformation company, is boosting its customer service by suggesting better responses and automating insights. And employee agents can work across people and teams to complete a ta task. One of the nation's leading health care providers, HCA Healthcare, is using our Generative AI technology to help caregivers spend more time with patients and less on pape paperwork. They're testing Katie. So, Katie is a Nurse Handoff Digital Assistant, helping to ensure continuity of care when one caregiver shift ends and another begins. It delivers summaries in everyday language to ensure that caregivers can share information easily and provides them up-to-the-minute patient data to expedite the discharge process. It also helps prepare patients for recovery at home by simplifying post-discharge tasks, like post-op physical therapy and more. Victoria's Secret & Company is looking to help in-store associates offer tailored recommendations to customers. The Home Depot's app called Sidekick, helps store associates manage inventory and keeps shelves stocked. Etsy uses Vertex AI to improve search, provide more personalized recommendations to buyers, optimize their ad models, and increase the accuracy of delivery dates. Vertex AI, especially when combined with Workspace, lets you create powerful employee agents. First, you create a custom model in the ways that we've shown before. From there, you connect them to all your company and web data. This can also be done with translation so that your company information is available, regardless of language. Similarly, we support multimodal inputs, including videos, call audio, images, in addition to text. Now, you will want to ground that in enterprise truth using databases like AlloyDB, BigQuery, and data from enterprise apps, like SAP, and announcing today, HubSpot. Let's take a look at an example of an employee agent in action. Please welcome Developer Advocate Gabe Weiss. >> GABE WEISS: Thanks, Aparna. Hi, folks! So, I know you all want to hear about awesome AI stuff that's coming, but I need to talk to you for a minute about my Annual Benefits Enrollment. See, I forgot, I have to finish signing up by today. And as you can see, I might be a little bit busy. So, if you don't mind, let's go ahead and look at this Open Enrollment email together. Okay, yep, I've got a deadline. I knew that. Thank you. I've got FSA stuff. I've got an online portal for my company, H R benefits. Okay, there's a lot here. Ooh, they included a video. Let's see if this makes my life easier. And the video is -- ah, okay. So, it's almost an hour long. Yeah, I'm not going to have time to review all of this stuff. Let's see how this Employee Agent that we've developed using Google Workspace, Gemini models, and Vertex AI might be able to help me. As you can see, it's integrated directly into my Google Chat, so I don't have to context switch while figuring this out. First things first, let's have it summarize the email and the video that it sent me. "Summarize the body and attached video from my recent email with subject: Open Enrollment Closing." So, behind the scenes, the agent is referencing that email body and its attachments as context in the prompt, using retrieval augment generation. That way, its response is limited to the content that matters to me. The Gemini model's multimodal capabilities allows the agent to understand and reason across text, audio, and video from a single prompt. I mean, this is a way quicker read. Okay, good. And I can immediately see that the medical plans have been completely revamped this year. Let's jump into the benefits portal to see more. Now, I've already done my dental and my vision, but I procrasti -- I mean, saved the most important plan for last, my medical plan. Let's see how this option stacks against my ex existing coverage. "Compare these coverage options to the PDF doc I have on the Platinum plan." The Gemini model's long context window, compared with Vertex Extensions enables the agent to cross-reference large amounts of data from a variety of sources, including unstructured data, like PDFs. Leveraging Gemini's advanced reasoning capabilities, the agent is able to understand the complex details of my current plan and compare it with the new options for 2025. And since the enterprise grounding features links me to the exact data that Gemini used to draw its conclusions -- which you can see linked here -- I can confidently trust its recommendation that the Gold plan is best for me. And done. So, now let's get a summary of my coverage. Let's say my house is multilingual, so I'd like to have it in Japanese also. "Please generate a summary of 2025 benefits in a Google Doc in both English and Japanese." Although my source material is in English, the Gemini model's support for over 40 languages enables it to understand and respond in Japanese. And here we go! (Applause) By engaging with employees in their native language, the agent can help everyone feel comfortable navigating their health care needs for the next year. So, one last thing. Now that I've officially completed enrollment, my daughter's going to need braces this year. Let's use the Agent to help me find the right orthodontist. "My daughter will need braces this year. Help me find the best in-network Orth donut orthodontists near my home." With oveVertex Extensions, the Gemini model can integrate with any external or internal API. This allows the Agent to join my dental coverage details directly with realtime Google Maps and Places data, helping to determine the best-ranked, in-network providers near me. Taking a quick look at my options. Okay, it looks like Cymbal Orthodontics is the only one that offers Saturday appointments, which works best for my family's schedule, but what do I really know about them? How do I know my agent actually picked the top orthodontists? Because the Gemini model natively supports grounding with Google Search, I can see what the Agent found and summarized, saving me hours of scouring the Internet for fresh information. The best part is I can get the annotations to see where the information came from, which means if I want to look for myself, I can. I have great places to start. And a quick scan of the summary shows me that they're great with kids, which is great. I particularly love this customer review. I'll go with them. Let's schedule a call. I don't know how it's going to find a schedule, though. I mean, I'm here. The agent knows that I'm at Google Cloud Next because it's integrated with Google Calendar, and it can find a time to schedule the call. Yes, please. Awesome! I've got a call. Do I need anything el else? Nope, I'm good. No, thanks. Perfect. Nice and easy. Just imagine what else an Employee Agent can help you with. Thanks! Back to you, Aparna. (Applause) >> APARNA PAPPU: That was great. Next up, Creative Agents. Generative AI is transforming how marketing teams are creating content, personalizing it, and making realtime campaign adjustments. Creative Agents can supercharge your design and production teams, working across images, slides, and other modalities. We're seeing really interesting Creative Agents from customers. For example, Carrefour, one of the world's leading food retailers, is pioneering new ways to use Generative AI for marketing. Now, using Vertex AI, they were able to deploy Carrefour Marketing Studio in just five weeks, an innovative solution to streamline the process of creating dynamic campaigns across various social networks, empowering their marketing teams to efficiently promote products and engage with their target audience. In just a few clicks, they can build ultra-personalized campaigns to deliver customers advertising that they care about. Other customers, such as Canva, are using Vertex AI to power its Magic Design for video, helping users skip tedious editing steps. Procter & Gamble is accelerating the development of ph photo-realistic images and creative assets, giving teams more time back to focus on higher-level plans. WPP, the creative transformation company, built a performance brain using Gemini 1.5 Pro on Vertex AI to power its media activation tool called WPP Open Media Studio. Advertisers, like the Google Pixel team, can now see predictions of which ads are going to work the best before a single impression has been served, avoiding millions of dollars of media waste over the life of campaigns. Now, while these examples are about professional creative teams, we believe that everyone can be a great creator and a great storyteller. But the formats and tools for storytelling at work haven't really changed that much. How many times have you heard, "Should we start with a doc or a deck?" Well, we can do a lot bet better. I am absolutely thrilled to announce our newest workspace app, Google Vids. (Applause) Fitting alongside Google Docs, Sheets, Slides, Google Vids is an AI-powered video creation app for work. With Gemini and Vids, you have a video, writing, production, and editing assistant all in one. Let me show you how simple it is to get started with Vids. Now, after a full week with all of you here at Next, I'm going to want to share a recap video to share all the excitement with my organization. When I open up Vids, Gemini helps me get started. I simply type in a prompt using an existing document for context. Now, based on that prompt, Gemini suggests a narrative outline for the story that I could easily customize and edit. I choose an expressive style, and Vids works its magic. Let's see what I get. So, wow! Just like that, I get the first draft with beautifully designed, fully animated scenes, complete with relevant stock media and music, and even a generated script. I'll be able to add fun videos and photos that I captured this week from my media, directly by accessing Google Drive and Google Photos, and even without any video expertise, I can build out scenes like a pro, just as easily as I make slides. But it doesn't have to stop there. With Vids, you can personalize your story your way by recording yourselves or by using one of our preset voices by Gemini, and by enriching scenes with high-quality media from our stock content library, or from your own, to really make your story shine. And because Google Vids is a Workspace app, it comes with all of the collaboration capabilities that your teams already know and love. Vids doesn't just help you create videos; it helps you become a better storyteller with a rich, new medium at work. This isn't just a vision; Vids is already in the hands of alpha customers who have been giving us really great feedback as we work towards expanding to Workspace Labs in June. I'm looking forward to sharing more about Vids with all of you in my Spotlight session tomorrow. Beyond Vids, many customers are building agents for their marketing teams, audio and video production teams, and all of the creative people out there who could use a hand. We provide a tremendously powerful platform and stack to build Creative Agents. First, it takes the best image generation model, Imagen 2.0, our most-advanced text-to-image technology helps businesses create images that match their specific brand requirements. This is now generally available in Vertex AI. Imagine strong language comprehension and photo realism capabilities contributed to Google securing the leadership position in the recent Forrester Computer Vision Wave. Second, as a part of Imagen, we're now introducing Text-to-Live Image, in preview today. Marketing and creative teams can generate animated images from a text prompt, including product images, ads, GIFs, and storyboards. Here's an example of a daylily opening in the morning dew and a pot of something delicious simmering on the stove, or an aerial view of a mountain range. Third, we are proud to announce the general availability of Digital Watermarking for AI-generated images produced by Imagen. It's powered by Google DeepMind's SynthID. Finally, we're announcing new editing modes for Imagen 2.0, which will make it really easy to remove unwanted elements in an image, add new elements, and expand the borders of the image to create a wider view for all of you. When all of these tools come together, it's a powerful combination. Let's take a look at the Creative Agents that you can build today. Please welcome one of our field engineering leaders, Eliot Danner. >> ELIOT DANNER: Thank you, Aparna. Our Cymbal Outfitters Product Team is working on designing a new tent that we're eager to bring to market. The tent itself is still in development, but we want to get started on a marketing strategy right away. Let's see how the Creative Agent we developed with Gemini models, Imagen 2.0's API, and Google Workspace can help. First things first, let's get a sense for the look and feel of our brand. "What are the key styles and themes of our brand?" The Creative Agent can analyze previous campaigns to understand our unique brand style and apply it to new ideas. In this case, the Creative Agent has analyzed over 3,000 brand images, descriptions, videos, and documents of other products that we have in our catalog contained within Google Drive to create this summary. And as you can see, it summarized our brand identity as bold, adventurous, and aspirational. This is exactly the look and feel that we're going for. The Creative Agent was able to use Gemini Pro's 1 million-token context window and its ability to reason across text, images, and video to generate this summary. Next, I'm going to prompt our Creative Agent using Imagen to dynamically generate amazing, new images of our tents in iconic and real outdoor settings. This will help us brainstorm campaign locations. All of this, of course, will match our unique brand style. Now, because I'm a terrible typist, I'm going to use Chrome's built-in clipboard history. So, we've asked it to create these images. We'll give it just a second. (Laughing) Check this out. Machu Picchu. Not bad. Half Dome. These are real landmarks, accurately presented. Now, let's work on a narrative to really help us land the message. I want to make special note that in this prompt, I'm going to ask the Creative Agent to provide captions in both English and Spanish, and that I'm using a multilingual prompt to do s so. Awesome. The Creative Agent leveraged Gemini Pro's multimodal capabilities to brainstorm creative captions that match each photo. These look great, too. And, thanks to our multilingual capability, we're able to share the same story with our customers in their native language. Moving along. Let's ask the Agent to combine everything into a storyboard and share it with the team to get some feed feedback. Generate a storyboard. Share it with an email group. The Creative Agent's going to move this into Slides and share it out with that group. All done. Shared with the group. They can post comments directly in Google Slides, and we can see them here for the Agent to review. We'll speed up time a little bit because it's a demo, and take a look at the feedback. Looks great, but there are three requests: First, something for less-experienced campers; second, content for the Cymbal Outfitters Podcast; and third, something more visually dynamic for social media. Let's start with that new storyboard. We'll use that copy/paste trick again, and you'll see in this case, we're asking for a new image, new captions, and a new storyboard with that new audience, that relaxed camping audience. Take a look. That looks pretty relaxing to me. Looks pretty good. And there's the story storyboard. Great! Just like that, the Creative Agent modified our ideas for a new audience; it understood our context and made the changes that we asked for. Now, let's generate that podcast. The Gemini model's complex reasoning allows me to generate a script and an audio clip in a single prompt. This much longer prompt highlights our ability to do ordered operations. We are asking the Agent to create a podcast script in SSML, or speech synthesis mark-up language, and then passing the output to our text-to-speech API. The result will be an episode with two people discussing tents. So, we'll see the script here, which we could expand, if we wanted, but let's just listen to a little bit of it. >> Welcome to the show. >> Thanks, Freya. It's great to be here. I'm always happy to geek out about gear. >> ELIOT DANNER: That sounds really good. And I want you to notice the regionalisms in the speakers' voices. Pretty impressive, but we don't have time to listen to the whole thing because you're here for the keynote. Finally, we were asked for some engaging content for social media. Let's see what the Creative Agent can do using our new, live image generation capability. So, we'll ask i it: "Generate some live images of our tent outdoors." For clarity, we've run this prompt before, and so, you'll be seeing cached to speed things up. But these are simply beautiful. Creating content like this could take days or weeks of scouting and shooting, but our Creative Agent used Imagen's new Text-to-Live Image capability to make them in minutes. The Creative Agent shows how Google's Generative AI capability can make producing custom content easy, seamless, and a lot of fun. I'll hand it over to Brad Calder. ♪ >> BRAD CALDER: Thanks, Aparna and Eliot! That was amazing. I want to first talk to you today about Data Agents. Many of us have worked with data analysts who can answer a question and figure out the questions we should be asking. That's what great data agents will provide. And one company that we're working with who's building data agents grounded in enterprise data is Walmart. Let's hear from Suresh Kumar, their EVP and CTO. ♪ >> SURESH KUMAR: Hello. And thank you for having me. At Walmart, we are obsessed with data. As the largest retailer in the world, the scale of our data is unmatched. Walmart has a legacy of innovating on behalf of our customers, our members, our associates, and today, our business is both vast and complex. We are working with massive amounts of complex data, and we need partners like Google Cloud to help us achieve our goals. The ability to handle huge amounts of data through BigQuery and other analytical tools has been instrumental in implementing our transformation. Artificial intelligence is embedded across our business, helping us unlock insights from our data to personalize experiences, get inventory to our supply chain, and unlock more. Using Gemini, we have enriched our data, helping us improve millions of product listings across our site, and ultimately, enabling customers to make better decisions when they shop with Walmart. Our team worked closely with Google to customize their offerings, fine-tuning for our needs to power the present and enable the future of Walmart. I'm excited about our partnership as we create the future of retail. (Applause) >> BRAD CALDER: Thank you, Suresh. We're thrilled to be helping Walmart use data agents to serve their customers. We're helping many companies build data agents. For example, Bayer is harnessing BigQuery and Vertex AI to develop digital medical solutions and drugs more efficiently, as well as streamline the creation of regulatory documentation. News Corp. Is using Vertex AI to search data across 30,000 so sources and 2.5 billion news articles updated daily. Woven by Toyota is collaborating with Google to leverage vast amounts of data in AI, resulting in 50% TCO savings to support automated driving. Mercado Libre is using BigQuery and Looker to optimize capacity planning and reservations with delivery carriers and airlines to fulfill shipments faster. And thousands of scientific researchers at Mayo Clinic search ed over 50 p petabytes of clinical data using Vertex AI Search. Tokopedia, an Indonesian e Commerce leader, is using Vertex AI to improve data quality, increasing unique products sold by 5%. And Vodafone uses Vertex AI to allow more than 800 operators to rapidly search over 10,000 contracts. Then, we have partners like Onyx and Data meMedica, who hel our customers build data agents. Now, these data agents have the potential to unlock new ways to find and act on meaningful signals from your data. They're built with AI for complex analysis, and they have to ensure factual integrity of their results. And to help power these data agents, we're making it easy to bring your data to Google Cloud with Gemini and BigQuery. Gemini and BigQuery helps with data preparation, discovery analysis and governance, and comes with a new BigQuery Data Canvas, which is a new notebook-like experience supporting natural-language queries and embedded visualizations. Now, data professionals can get assistance for building their data pipelines to improve performance, reduce errors, and lower costs. And today, we're introducing new query capabilities using Vertex indexing directly in BigQuery and AlloyDB. This allows you to leverage AI over your data where it is stored, enabling realtime and accurate responses for AI applications. We're also building seamless connection from BigQuery to Vertex AI to give you direct access to AI models in Vertex from BigQuery. Now, BigQuery and Vertex together deliver up to four times faster performance and are up to three times more cost-effective than the other data warehouse and AI platforms. Now, this allows you to seamlessly combine your AI models, classical ML operators, and statistical functions together in your data pipelines for multimodal analysis. And with this new capability, you can analyze documents, videos, images, and audio recordings together with your structured da data. And finally, we've improved the integration between Looker and your data agents as part of the Gemini and Looker Preview. We've added new AI capabilities that enable you to chat with your business data with seamless integration with Google Workspace. And together, these solutions make it easier to analyze and activate your d data. Let's see how this comes together in a demo. Please welcome Director of Outbound Product Management in Data Analytics Yasmeen Ahmad. >> YASMEEN AHMAD: Thank you, Brad. We have built a great data agent, one that connects all of your data, incorporates gen AI, and even has collaboration. It's a custom application built with BigQuery, Looker, and Vertex AI. (Notification sound) Shoot! I could have sworn my notifications were on silent. But wait, what's this? This is cool. It's my data agent, giving me a notification that sales are up. Well, maybe it's because of our Cloud shoes. Because surely, shouldn't all keynote speakers get keynote sneakers? (Laughter and applause) All right, let's explore with our data agent. What you're seeing here are realtime KPIs, the metrics that matter the most, being continuously monitored from my enterprise data in BigQuery to surface anomalies and trends. Let's dig i in. Let's explore Cloud Shoe sales over customer segments and territories in a heat map. Our data agent uses the Looker
