Video Summary
google presents an 'agentic enterprise' blueprint to move AI from pilots to production with a unified stack
gemini enterprise agent platform enables low-code agent building, governance, observability, and a partner marketplace
infrastructure advances: ai hypercomputer and eighth‑generation tpus to accelerate training and serving
agentic data cloud (knowledge catalog, cross‑cloud lakehouse, lightning engine) unifies and makes data agent-ready
agentic defense introduces autonomous security operations and partnerships (e.g., Wiz) to detect and remediate threats at machine speed
The agentic enterprise is Google Cloud's blueprint for moving AI from pilots to organization‑wide production by using a unified stack—aligned chips, models, data, agents and secure infrastructure—to enable autonomous, agent-driven workflows at scale.
Gemini Enterprise offers a full‑stack agent platform with low‑code agent building, an agent registry, skills/tools packages, a partner marketplace, agent identity and gateway for zero‑trust security, and observability/optimization to manage agents lifecycle and governance.
Google announced large CapEx investments and the AI Hypercomputer with purpose‑built infrastructure including eighth‑generation TPUs, designed to accelerate both training and serving for large models and to improve price‑performance for heavy workloads.
Security is handled via Agentic Defense: an autonomous SOC that uses agents to triage and remediate threats at machine speed, integrated with partners like Wiz which build dynamic inventories, security graphs, and automated remediation workflows.
"Today, that future is running in production at a scale that the world has never seen."
Thomas Kurian, CEO of Google Cloud, emphasizes the significant progress made in AI over the past year, highlighting that nearly 75% of Google Cloud customers are now utilizing AI products to enhance their operations.
There has been a shift from experimentation to widespread implementation of AI in business settings, which presents a new challenge: integrating AI into production across entire organizations.
"To drive real value, you need an architecture where chips are designed for the models."
Kurian stresses the necessity of a unified stack for AI deployment, arguing that fragmented systems cannot deliver the desired outcomes.
He explains that Google’s own AI systems, which power services like Search, YouTube, Chrome, and Android, rely on an integrated architecture that aligns chips, models, data, and applications.
"This year, we plan to invest between $175 to $185 billion in total CapEx, a nearly 6x increase in just four years."
Sundar Pichai, CEO of Google, announces substantial investments aimed at advancing AI and cloud capabilities, particularly in machine learning.
The goal is to ensure Google Cloud maintains its position at the forefront of technology, which will, in turn, benefit all customers through enhanced AI functionalities.
"We created a system of agents taking on three different types of roles—planners, orchestrators, and coders."
Pichai details how Google has embedded AI into internal workflows, especially in code generation, where 75% of new code is now AI-generated, up from 50% last year.
He describes a recent code migration project that utilized AI agents to significantly expedite the process, completing it six times faster than previously possible.
"Teams used our models to rapidly generate thousands of variations of our creative assets."
The marketing teams at Google leverage AI for creative tasks, leading to faster campaign adaptations and increased conversion rates, achieving a 70% faster turnaround.
In security, AI agents have automated the triage of threat reports, cutting threat mitigation times by over 90%, showcasing AI's ability to enhance operational efficiency in critical areas.
"Today, we're taking the next step in bringing Google AI to every employee and every workflow."
Kurian introduces Gemini Enterprise as a comprehensive system designed to connect data, people, and business objectives, thus facilitating the implementation of AI across all workflows.
The initiative aims to create a cohesive environment that enhances both intelligence and automation, essential for achieving effective business transformations.
"Today, we unveiled Citi Sky, an always-on AI-powered member of the Citi Wealth team that brings Citi's global intelligence to clients' fingertips."
Citi Sky is introduced as a revolutionary AI member intended to enhance client interactions within the Citi Wealth division.
This AI solution offers an exemplary experience for clients and colleagues, facilitating inquiries in multiple languages at any time.
Honeywell is generating billions of insights for building management through digital twins trained on over a million product specifications, showcasing the scale of AI applications in various industries.
"Liverpool is bringing their signature in-store service online, projecting a ten times return on investment on their new shopping assistant."
Retail innovation is exemplified by Liverpool's move to digitize their in-store customer service, leveraging technology to improve online user experiences.
They anticipate a substantial return on investment through the implementation of their shopping assistant, which is designed to mimic the quality of in-person service.
"We were proud to partner with NASA to use Gemini Enterprise Agents to power flight readiness and ensure astronaut safety for Artemis II."
The partnership with NASA illustrates the profound reach of AI, particularly in mission-critical areas like spaceflight.
Gemini Enterprise Agents are being utilized to enhance safety protocols for Artemis II, marking significant advancements in human spaceflight capabilities.
"We built the agent platform to manage the entire lifecycle of an agent."
The agent platform incorporates low-code solutions, allowing every employee to create and deploy agents through natural language interfaces.
This framework ensures the integration of specific business rules into large language models, resulting in predictable and autonomous actions across workflows at scale.
"The agent registry provides a single point of control, indexing every internal agent and tool across your organization."
A robust management system is established with the agent registry, facilitating the discoverability and governance of internal tools and agents within an organization.
The Skills and Tools Registry allows companies to define reusable packages of instructions, scripts, and resources, enhancing efficiency in training agents for specific tasks.
"Our Agent Marketplace allows you to search and deploy specialized agents from our global partner ecosystem."
The Agent Marketplace enhances accessibility to specialized agents from various partners, ensuring that organizations can easily integrate external expertise into their operations.
Major partners like Atlassian, Box, and Oracle are part of the ecosystem, providing a diverse array of tools for many business needs.
"With Agent Identity, every agent has a unique cryptographic ID and well-defined authorization policies."
The platform implements zero-trust security principles, ensuring that every action taken by agents is transparent and traceable, essential for maintaining security within an organization.
The Agent Gateway serves as a centralized management solution for enforcing policies, creating a command center that enhances administrative efficiency.
"Agent observability delivers granular instrumentation, allowing you to visualize the full execution path of any agent."
The integration of optimization and observability features provides extensive monitoring capabilities, enabling organizations to diagnose and assess their agents’ performance with precision.
Instrumentation aids in identifying reasoning loops and promoting proactive resolutions, which is crucial for maintaining operational efficiency.
"Gemini Enterprise is the primary environment where your business actually operates."
Serving as a simplified interface for interacting with advanced AI capabilities, Gemini Enterprise enables employees to leverage technology without needing extensive technical knowledge.
By unifying data across various platforms, Gemini Enterprise creates an intelligent flow that enhances collaborative efforts within organizations.
"Gemini Enterprise is a front door for AI in our business."
Project 3D is part of Gemini Enterprise's initiative to empower employees, particularly in customer-facing roles, enhancing confidence in responses through natural language intelligence.
Virgin Voyages shares how integrating Gemini Enterprise into their operations has dramatically improved efficiency and service delivery, resulting in substantial increases in sales.
"Gemini Enterprise is a platform to build, manage, and interact with agents."
A demonstration showcases the capabilities of Gemini Enterprise to orchestrate multiple AI agents, optimizing workflows and improving business outcomes effectively.
Users can engage with a personalized homepage that integrates various internal and external business contexts, enhancing decision-making processes.
"With that one prompt, multiple agents complete a series of actions in minutes instead of hours."
The video illustrates how agents quickly analyze market trends, manage inventory, and suggest strategies, showcasing the efficiency gained through automation.
By leveraging insights from various data sources, the Gemini Enterprise platform enables organizations to respond dynamically to market changes, exemplified by a proposed relaunch campaign for dormant products.
"Now, let's get that website updated."
The process begins with coordinating tasks among team members. The development agent is tasked to work with the engineering team to update the website effectively.
Integration with Jira is highlighted as essential, providing developers full context for seamless task handoffs. Notifications for any updates or changes appear in Google Chat, ensuring transparent and efficient communication.
Using the Command Line Interface (CLI), the developer agent starts working on the specific Jira ticket, which includes an overview of strategies, assets, and execution plans based on previous sessions and guidelines.
"I gave it a little bit of a head start."
Deployment can be time-consuming, so a head start is initiated to keep the process moving smoothly. After some time, they review the final product and find it to meet their expectations.
Moving back to Gemini Enterprise, preparations for the store launch are vital. The store operations team typically handles these tasks, and the speaker starts a new session to gather necessary information.
Gemini Enterprise leverages the company's context to create customized content like decks for regional distributors, tailoring it to reflect past management experiences and sales goals.
"With our brand-new Canvas mode, I can make edits and collaborate in Google Slides."
The new features allow users to edit and collaborate within Google Slides without leaving the Gemini Enterprise platform, streamlining the workflow.
The speaker shares the deck with their team, facilitating immediate engagement and collaborative efforts on the content.
This setup reflects a seamless blend of context-sharing and content generation, with embedded governance ensuring secure management at scale.
"Thousands of companies around the world and across industries are choosing Gemini Enterprise to transform their business."
Many organizations are not just redesigning workflows but empowering employees to become AI builders, drastically changing how they approach problem-solving.
Case studies from various enterprises show significant adoption rates within weeks leading to improved efficiencies and user engagement, demonstrating the platform's effectiveness.
Specific examples include Signal Iduna in Germany, Bosch, KPMG, the American Society for Clinical Oncology, and Merck, highlighting the widespread impact and adoption across industries.
"At Walmart, we're known for being people-led and tech-powered."
Walmart leverages technology to strengthen relationships with associates and customers by equipping employees with advanced tools that keep them present on the sales floor.
The integration of Google Cloud and Gemini Enterprise is enabling Walmart to make data-driven decisions faster, enhancing customer service and operational effectiveness.
With leaders now equipped with real-time insights via connected devices, the emphasis is on creating an environment that fosters trust and improves service delivery.
"Google Cloud is helping athletes push the boundaries of their sport with AI."
Google Cloud is collaborating with athletes to enhance performance through data analytics and AI. This initiative aims to uncover critical insights that can give athletes a competitive edge.
The focus is not only on current performance but also on analyzing past performances to draw comparisons and identify areas for improvement.
Embracing cutting-edge technology allows athletes to adapt and excel in their respective sports, showcasing the versatility of Google Cloud beyond typical business applications.
"That data is really going to help me progress."
The ability to analyze data related to athletic performance is crucial for progression in sports. It allows athletes to compare their performance metrics and identify areas for improvement.
The next generation of athletes will similarly leverage these tools to enhance their skills and understanding of their sports.
"This really pinpoints it exactly."
Innovations in technology enable detailed visualization of complex athletic maneuvers, such as Shaun White's snowboarding tricks.
This increased visibility can enhance the viewing experience for fans, as they can better appreciate the skill and intricacies involved in the sport.
"Google Cloud is helping everyone innovate here."
Google Cloud plays a pivotal role in advancing technology in sports by developing specialized models for data analysis, spatial assessment, and secure data management.
The infrastructure provided by Google Cloud, including powerful TPUs (Tensor Processing Units), supports athletes and enterprises in their quest for innovation.
"We're taking the exact same technology that empowers athletes and applying it to the enterprise."
The same data-driven technologies that assist athletes in understanding their performance are now being utilized to revolutionize enterprise operations.
This transformation emphasizes the need for scalable systems capable of processing vast amounts of data to enhance organizational efficiencies.
"I am proud to announce our eighth-generation TPUs."
The latest TPUs represent significant advancements, with specialized platforms engineered for the distinct demands of training and serving in AI applications.
Innovations include improved performance capabilities, reducing training times dramatically and allowing for the analysis of larger datasets than ever before.
"We've partnered with Google Cloud to build what we think is the preeminent scalable cloud-based research environment."
Citadel Securities, a leading financial institution, has partnered with Google Cloud to optimize their operations, achieving faster processing times and reduced costs.
Utilizing TPU technology allows them to run complex workloads with enhanced efficiency, significantly benefiting their trading and research capabilities.
"Imagine you're a quantitative researcher with 100 novel ideas. Can we empower our researchers to test all their ideas, limited only by their creativity?"
The discussion highlights the shift in executing research ideas from weeks or days to mere hours or even minutes.
It emphasizes the vision of enabling researchers to focus on their creativity rather than being restricted by platform capabilities or economic factors.
"At this scale, humans cannot manually troubleshoot configurations; you need a cloud that drives itself."
Google Cloud's collaboration with Citadel Securities serves as a prime example of leveraging foundational AI Hypercomputer technology for competitive advantage.
The integration of the Model Context Protocol allows for automated orchestration of Google Cloud services, enhancing efficiency in handling configurations.
"Today, we are introducing the Knowledge Catalog, a universal context engine for your enterprise."
The Knowledge Catalog is designed to unify structured and unstructured data, facilitating seamless integration with existing platforms like BigQuery.
It autonomously processes files, enriching data and making it agent-ready without the need for manual data engineering.
"This trusted context enables agents to act with certainty."
The Knowledge Catalog integrates with Gemini Enterprise's Deep Research Agent to provide multi-step reasoning for accurate business insights, drastically reducing manual effort.
This significantly accelerates the decision-making process compared to traditional methods that may take weeks.
"We are unveiling a Gemini-powered data science authoring experience that transforms how data practitioners work."
The Data Agent Kit is introduced as a tool that integrates with existing workflows, allowing practitioners to autonomously build data pipelines and deploy models efficiently.
This transformation frees researchers to concentrate on innovation rather than manual data management.
"Lightning Engine delivers up to 2x the price performance over the previous market leader."
The Lightning Engine for Apache Spark is developed to meet the immense demands of current data processing requirements, enhancing performance and efficiency.
The platform serves industry leaders, such as Flipkart and Lowe's, to accelerate their workload processing.
"Today, we are introducing the Cross-Cloud Lakehouse, where your analytical engine reasons over data in any cloud."
This new architecture allows for low-latency connectivity to various cloud environments, removing the barriers of complex networking and high egress fees.
The focus is on providing businesses with freedom from vendor lock-in, facilitating seamless data access across multiple platforms.
"Vodafone unified their data on BigQuery to build a more resilient global network for their customers."
Companies like Vodafone and Macquarie Bank demonstrate the transformative power of the Agentic Data Cloud, using it to enhance operational efficiency and customer experience.
By centralizing data and deploying AI-driven solutions, these organizations are able to optimize operations significantly, illustrating the platform's practical impact on business.
"Midnight Swirl contains soy, and it's giving me a data citation, a schema called product specs."
The presentation begins with a focus on utilizing "dark data" to extract valuable insights, particularly relating to ingredients and allergens in products.
A schema known as "product specs" has been developed using the Knowledge Catalog to analyze supplier and recipe PDFs, allowing for the extraction of critical entities such as recipes, ingredients, and allergens.
This process uncovers previously hidden connections among these entities, providing trusted context for decision-making.
"With our cross-cloud Lakehouse, the data stays exactly where it is while we dynamically build the list."
The speaker explains the challenge of accessing distinct datasets, in this case, allergen schemas in Google Cloud and customer loyalty lists in AWS S3 Iceberg.
Instead of undergoing a lengthy migration process, the cross-cloud Lakehouse enables dynamic list creation without relocating the data, thus allowing for increased efficiency.
Gemini, the data engineering tool, generates a plan while respecting existing security permissions, making collaboration easier for data scientists.
"With our Lightning Engine, filtering our soy-sensitive customers happens in mere seconds."
The execution of tasks is accelerated by using the Lightning Engine, which significantly enhances price-performance compared to competing alternatives.
The process demonstrates the speed at which customer lists can be filtered based on specific dietary sensitivities, illustrating the effectiveness of using AI technology in real-time data processing.
The ability to generate an annual forecast helps to determine whether a global launch is feasible, showcasing a sophisticated approach to decision-making.
"In an agentic enterprise, giving agents the power to act means that security must become an autonomous force, moving faster than the threats themselves."
The keynote addresses the urgent need for security measures that can respond more quickly than evolving threats, particularly in an era of AI-driven attacks.
Human analysts are becoming slower in comparison to automated systems, as vulnerabilities can be exploited before patches are developed.
Google Cloud's new agentic Security Operations Center (SOC) leverages automation to significantly reduce investigation times from thirty minutes to just sixty seconds.
"To protect this new frontier, security must now move at machine speed."
Wiz co-founder introduces their approach to AI security, emphasizing that traditional security measures must adapt to the rapid pace of threats in the AI landscape.
The Wiz platform aims to provide deep visibility into AI applications, enabling proactive risk management and empowering development teams to implement secure practices from the start.
Key innovations include the use of AI agents to autonomously identify and remediate security risks, significantly enhancing the security posture of enterprises.
"Wiz starts by automatically building a dynamic inventory from your code and cloud, completely agentlessly."
The Wiz platform automatically constructs a comprehensive inventory of an organization's code and cloud resources to facilitate better visibility and security.
By generating a "security graph," Wiz creates a living map of AI application architecture, helping to identify potential vulnerabilities efficiently.
The use of automated agents streamlines processes like triage, suggesting fixes and unearthing issues without the need for cumbersome manual meetings or spreadsheets.
"AI agents unlock secure innovation with AI at scale."
The emergence of AI agents significantly enhances security within organizations, enabling security teams to perform crucial tasks quickly and efficiently.
These agents can identify their architecture in real time, conduct security reviews rapidly to assess risks, and suggest fixes, all while integrating into developers' native tools.
This streamlined process fundamentally changes how security measures are implemented, allowing teams to focus on innovation without compromising security.
"The unique combination of Google's agentic SOC and Wiz's deep cloud context gives customers the confidence to build and deploy generative AI everywhere safely."
Collaborations between Google Cloud and Wiz are instilling confidence in organizations by providing advanced security measures that protect critical infrastructure.
Notable partnerships include initiatives to secure essential systems for the Los Angeles Department of Water and Power and proactive defenses against cyber threats in Singapore.
Major financial institutions like Morgan Stanley are choosing Wiz to bolster visibility and control across their cloud environments, highlighting the growing need for robust cloud security strategies among leading organizations.
"A secure foundation gives you the confidence to innovate."
Organizations are leveraging a coordinated agentic task force, where specialized agents support teams rather than merely serving as tools.
This new paradigm involves the agents acting as strategic extensions of the business, enhancing engagement, expanding reach, and providing personalized service at scale.
Tools like Gemini Enterprise for Customer Experience are at the forefront of this innovation, reshaping customer interactions through seamless integration of AI across various channels.
"With pre-built shopping and food ordering agents, every interaction turns into revenue and retention opportunities."
The introduction of agents for commerce greatly enhances the customer journey from discovery to service interactions, creating a more cohesive experience.
Companies like Papa John's utilize food ordering agents that remember customer preferences to facilitate quicker service.
Best Buy employs Agent Assist to help employees provide accurate answers and enhance customer engagement, showcasing how AI can improve customer service operations effectively.
"YouTube TV recently launched a voice agent that serves 100% of their users live in production."
The deployment of a voice agent in YouTube TV's customer support exemplifies how AI can address complex queries while seamlessly managing multiple languages.
The CX Agent Studio allows for easy modifications without needing extensive technical updates, making it user-friendly for teams to enhance their service offerings.
This technology is crucial for improving customer interactions and ensuring quick, effective responses to inquiries across various platforms.
"With our built-in test interface, you can ensure that every answer is grounded, factual, and pulled directly from a knowledge base."
Google has developed multi-specialized agents that can efficiently manage even the most complex customer requests, ensuring accurate and factual responses sourced directly from an established knowledge base.
A practical example is demonstrated by YouTube TV, which successfully built and deployed this entire support experience in just six weeks.
The simplicity of adding specialized agents is showcased; for instance, introducing a new promotions agent can be done swiftly with a few clicks, further integrating business logic in natural language into the agent's capability.
"We just turned a routine support call into a multilingual upgrade and a globally deployed promotion, all in under four minutes."
The effectiveness of the multi-agent system is illustrated by transforming a simple support call into a multilingual service and deploying promotional activities globally within a very short timeframe.
This efficiency highlights the potential of the Gemini Enterprise platform, which offers enterprise-level reasoning, robust tool integration, a powerful visual builder, and valuable insights for businesses to respond quickly to changing demands.
"What if you could skip both finding information and figuring out what to do with it?"
Google Workspace, recognized as the world's leading productivity tool, aims to combat the fragmentation of work by introducing Workspace Intelligence, which streamlines the process of gathering and utilizing information.
This innovation proposes a unified intelligence layer residing within every Workspace application, allowing artificial intelligence to perform the heavy lifting needed for efficient collaboration and organization.
"With Ask Gemini, Google Chat becomes my Command Center."
A demonstration shows how Workspace Intelligence acts as a command center in Google Chat, providing real-time updates and prioritizing urgent tasks while eliminating the need for multiple open tabs.
By leveraging advanced AI features, users can quickly access relevant documents, such as merchandising playbooks, efficiently linking information without manual searching.
"Workspace Intelligence is the end of the context tax."
Workspace Intelligence greatly simplifies the process of converting fragmented data into organized plans by allowing users to give commands in natural language, instigating automatic document creation.
This removes the burden of tedious data management, unveiling a seamless method of generating professional presentations that align with corporate branding and individual user formats—all executed without cumbersome setup procedures.
"With Google Cloud, we are turning intelligence into a growth engine."
The applications of Google Workspace and Gemini technology are vividly illustrated by various companies driving innovation and business growth through the platform.
Examples include Colgate-Palmolive, which streamlined operations for 34,000 employees, and Natura, which achieved a tenfold acceleration in data-driven reporting.
These instances demonstrate that AI not only enhances productivity but also transforms traditional workflows, enabling organizations to respond swiftly to market needs while maintaining operational excellence.
"Migrating your entire organization, including complex finance teams from Microsoft 365 to Google Workspace, is now up to five times faster."
The migration process to Google Workspace has significantly improved, allowing organizations to transition efficiently without being bogged down by complexity.
This enhanced migration capability is designed to support organizations eager to harness the advantages of Google’s workspace tools while minimizing disruption and enhancing productivity.
"We're bringing AI to small and medium-sized businesses by helping them adopt Gemini Enterprise and our AI advances in Workspace."
Google is focusing on integrating AI into small and medium-sized businesses through their Gemini Enterprise and Workspace advancements, showing a clear strategy to democratize AI technology.
The presentation highlights the introduction of a variety of new technologies designed to enhance business operations and improve customer experiences.
"We've introduced our agentic blueprint as a foundation for agents to transform your companies."
The agentic blueprint is presented as a transformative foundation for organizations aiming to leverage AI agents, suggesting that businesses must adapt to harness the full benefits of AI.
Key components include the AI Hypercomputer, Agentic Data Cloud, Agentic Defense, and the Gemini Enterprise Agent Platform. Each of these elements is essential to building a robust AI framework within companies.
"First, the AI Hypercomputer, the purpose-built foundation optimized for the physics of the agentic era."
The AI Hypercomputer serves as a specialized infrastructure tailored to support the demands of modern AI applications and workloads.
The Agentic Data Cloud acts as a vital engine, providing agents with the necessary trusted business context to operate effectively in real time.
Agentic Defense is discussed as a critical feature that ensures the security of the entire AI lifecycle, promoting trust in AI solutions.
"To each of our customers and partners, thank you for being on this amazing journey with us."
The speaker expresses gratitude towards customers and partners for their continued engagement, emphasizing the collaborative nature of this technological journey.
There is a strong message of pride regarding the hard work of teams that developed these innovative platforms, reinforcing their readiness for deployment.
"We have three fantastic days planned for you, with a great set of sessions and spotlights still to come."
Attendees can look forward to a series of enlightening sessions planned for the event, indicating that the conference will provide ample opportunity to learn about the new technologies and applications in depth.
This sets the stage for a robust discussion and exploration of the capabilities introduced during the keynote presentation.
