Build 2025: Microsoft’s AI Agent Revolution Meets Google’s Hyper-Personalized AI Strategies

[00:00:02] It's Wednesday, May 21st, 2025, and I'm Dave Sobel. We talk Build and I.O. today. Microsoft redefines the operating system as an AI agent platform embedding identity, protocol, and lifecycle tools. Google bets on hyper-personalized AI across every product layer. Both giants converge on the model context protocol, hinting at a new standard for agent interoperability. And subscription fragmentation deepens as Google launches its $250 AI Ultra.

[00:00:32] This is the Business of Tech.

[00:01:02] Microsoft's new features are available to the existing workflows. Key features unveiled include the public preview of Microsoft IntraAgent ID, which enhances identity and access governance for AI agents. Microsoft aims to provide organizations with the same protections for AI agents as they do for human employees. Microsoft announced the integration of Native Model Context Protocol, or MCP, into Windows, which aims to enhance the interaction between AI applications and the operating system.

[00:01:28] Microsoft's new features, described by Microsoft, and the OSBC port for AI applications, allows developers to create AI agents that can seamlessly connect with various applications and services, thereby automating user tasks more efficiently. The model context protocol was introduced by Anthropic and is expected to facilitate easier communication between AI agents and the Windows environment, enabling functionalities like natural file searches.

[00:01:54] Microsoft emphasizes security in this early phase, offering the MCP framework to select developers to address potential vulnerabilities such as token theft and server compromises. Microsoft announced the rollout of Microsoft's new features for Microsoft 365 co-pilot during build, focusing on enhancing human agent collaboration through innovative tools.

[00:02:15] Microsoft has also launched the Microsoft 365 co-pilot tools, where users can access the company's reasoning agents and third-party tools, and the

[00:02:45] Microsoft's new features for Microsoft's new features. Microsoft has been developed by Microsoft's new features, and the Microsoft's new features, and the Microsoft's new features, and the Microsoft's new features, and the Microsoft's new features. Microsoft unveiled a new open protocol called NLWeb, aimed at transforming how websites integrate artificial intelligence for search functionality. This innovation allows any website or app owner to deploy customized chatbots with minimal coding, enabling a more personalized and efficient search experience.

[00:03:23] And Microsoft is looking to maintain its leadership in the artificial intelligence coding space by open sourcing its GitHub co-pilot code extension and integrating into Visual Studio Code.

[00:03:39] With Visual Studio Code being the most widely used coding editor globally, Microsoft aims to counter competition from startups like Cursor and Windsurf, which have developed their own proprietary AI coding tools with new metrics and frameworks needed to assess productivity in the emerging era. Microsoft announced that Microsoft announced that it will host the Grok 3 AI models from XAI, founded by Elon Musk, expanding its Azure AI Foundry offerings.

[00:04:05] This addition could stir controversy within the company and heighten tensions with its partner, OpenAI. Microsoft confirmed that these models will be hosted and built directly by them, ensuring that they meet the service level agreements expected by Azure customers. This move aligns with Microsoft's strategy to position itself as the primary hosting provider for emerging AI models, reflecting its rapid growth in the Azure AI Foundry sector.

[00:04:28] And Microsoft announced that it is making the Windows subsystem for Linux open source, allowing the developer community to contribute to the code. This move follows nearly nine years since the initial launch of the feature, which enables users to run a Linux environment within Windows. Microsoft is doubling down on AI agents as core operating system primitives, signaling their intention to shift how users interact with software. This is not just about assistants like Copilot.

[00:04:57] It's about AI becoming a native layer in Windows with secure identities, system level access and protocol driven integration. Build 2025 isn't a Copilot update event. It's a declaration that the OS is becoming an agent platform. This matters to IT service providers because Microsoft is both paving and controlling the road ahead. These changes impact how customers interact with technology, which tools they need to adopt and what kind of help they'll need. AI agents aren't just productivity features.

[00:05:24] They're the new users in a managed environment and managing them is about to become a core competency. Microsoft is pushing toward a future where AI agents are first class citizens in the digital workplace. This requires redefinition of identity management, system integration and customer workflows. For providers, this creates both new complexity and new services. Agent lifecycle management or identity and policy frameworks for non-humans. Agent driven automation consulting or developer productivity enablement.

[00:05:54] Will it all happen? Maybe. Regardless, it's important to understand their intention. With every new breach and threat that I cover, it's clear that cybersecurity isn't a luxury anymore. It's a necessity. That's where Huntress comes in. Their fully managed cybersecurity platform is built for every kind of business. Not just the 1%. Huntress seamlessly integrates their products and threat hunting team.

[00:06:20] Their EDR, ITDR, SIM and security awareness training solutions are purposely built for their elite 24x7 security operations center to stop threats before anyone else even spots them. This potent combination of purpose-built cybersecurity and threat hunting expertise is one of the many reasons why G2 users have voted Huntress the number one rated EDR for growing businesses.

[00:06:46] To see what people-powered cybersecurity looks like, visit Huntress.com slash MSB Radio. And starting a day later, Google announced significant upgrades to its artificial intelligence capabilities at its annual I.O. conference, aiming to enhance user experience and fend off competition from emerging startups like OpenAI.

[00:07:07] The company introduced an AI Ultra Plan subscription model priced at $249.99 per month, which offers users advanced AI features, including access to experimental tools. Alphabet's chief executive officer, Sudharpa Chai, reported that the company's AI assistant, the Gemini app, has surpassed 400 million monthly active users. With the updates, Google is shifting its search engine to include an AI mode, designed to provide computer-generated answers for complex queries instead of traditional web links.

[00:07:36] Analysts suggest that Google's market share in search may drop below 50% in the next five years if current trends continue. Google's artificial intelligence models are gaining a competitive edge through the integration of personal data, enhancing their ability to deliver tailored responses. The company's latest feature, called Personal Context, allows the model, Gemini, to access user information from various Google applications, provided users get their permission.

[00:08:01] This development follows Google's introduction of the Gemini with personalization feature earlier this year, which utilizes users' search histories to create more relevant and insightful responses. According to Pachai, this capability enables the AI to generate replies that reflect users' communication styles by analyzing past emails and documents. Google announced its support for the model context protocol, or MCP, during I.O. and integrated it into the Google Gemini API and software development kit.

[00:08:31] MCP has rapidly gained traction as a standard protocol for linking artificial intelligence agents to various tools and data sources. According to Pachai, protocols like MCP and the Agent-to-Agent protocol are crucial for enhancing capabilities of AI agents, allowing them to connect more efficiently to data. And Google's vision for search is evolving as the company rolls out that AI mode aimed at enhancing user experience with a more conversational search interface.

[00:08:57] The feature, which allows users to engage in dialogue with the search engine and receive tailored responses, is part of a broader strategy to integrate artificial intelligence into its core products. The new AI mode uses a query fan-out technique, enabling multiple queries to be processed simultaneously, thereby delivering more comprehensive results. Additionally, Google plans to introduce a deep search feature that will provide users with detailed, cited reports in response to complex questions, as well as real-time data visualizations.

[00:09:27] Google unveiled its latest image generation model, Imogen 4, which the company claims features enhanced quality and improved spelling capabilities. According to Eli Collins, Vice President of Product at Google DeepMind, this model offers remarkable clarity in intricate details and excels in both photorealistic and abstract styles.

[00:09:47] Imogen 4 will be available starting May 20 across various platforms, including the Gemini app, Wisk, and Vertex AI, as well as in Google workspace applications like Slides and Docs. The new model is reported to be up to 10 times faster than its predecessor, Imogen 3, which could significantly enhance user experience in generating images for various applications such as greeting cards and comics. Google introduced the Gemini AI assistant to its Chrome browser.

[00:10:15] The assistant is designed to clarify complex information and summarize web content, reflecting that broader trend as businesses increasingly adopt AI. In addition, Google unveiled Veo 3, an advanced video generator that integrates artificial intelligence-generated audio, which competes directly with OpenAI's Sora. The company also launched Flow, an AI filmmaking tool that allows creators to manipulate shots and camera angles using generative technology.

[00:10:42] Furthermore, Google is enhancing its existing projects, including Project Astra and Project Mariner, to improve user interaction with real-time AI capabilities. Notably, the Synth ID detector tool was introduced to embed invisible watermarks in AI-generated content, ensuring transparency in the use of AI across various media. Over 10 billion pieces of content have already been watermarked using this technology.

[00:11:08] And Google introduced a new feature for Google Meet that enables near real-time translation of spoken language during video calls. This AI-powered functionality, revealed at I.O., allows users to communicate seamlessly with colleagues who speak different languages by translating speech while preserving the speaker's voice and tone. Initially, the feature supports translations between English and Spanish, with plans to expand to Italian, German, and Portuguese in the coming weeks.

[00:11:34] The service is available in beta for users subscribed to Google's AI Pro plan or the newly launched AI Ultra subscription, costing $249.99 per month. Why do we care? Google is going all in on hyper-personalized, multimodal, and deeply embedded AI. Not just across products like search and docs, but at the infrastructure and protocol layer.

[00:11:57] While Microsoft is anchoring its strategy in enterprise-grade identity and OS-level integration, Google is betting on the value of context and ubiquity, positioning Gemini as your personal digital shadow. Everywhere, everywhere, all the time, with access to everything, and hopefully following through on that consent. I didn't mention it before, but note both companies announcing support for MCP, a rare move toward standardization in AI agent architecture and quickly.

[00:12:25] If MCP is indeed becoming the USB-C of AI, this is the equivalent of the two biggest OS vendors agreeing on a universal plug format for autonomous systems. This matters for IT service providers not because of what MCP does today, but because of what it enables. Interoperable agents, portable context, and system-level AI abstraction. And what it requires. Governance and security frameworks.

[00:12:50] Expansion of personal context beyond Gemini aims to enhance user experiences across platforms like Search and YouTube, potentially positioning Google ahead of competitors such as OpenAI, whose models start with no prior knowledge of users. Google's vision is to create contextual, creative, and ubiquitous AI. While Microsoft's strategy focuses on infrastructure, Google's approach prioritizes users, and both will collide in customer environments.

[00:13:16] So for providers, privacy access and agent governance become new core competencies. New opportunities open around search content optimization, AI-enabled creative support, and Gemini co-pilot integration planning. And providers must help clients navigate licensing chaos across increasingly segmented in AI subscription offerings. This episode is supported by Comet Backup.

[00:13:42] Whether you get hit with ransomware, hardware failure, or human error, there's nothing more heart-stopping than losing business-critical data. Backups are your final stand when a threat penetrates your layers of defense. That's where Comet Backup comes in. Comet is an all-in-one backup solution. Whether you need to protect computers, servers, virtual environments, emails, or databases, Comet Backup empowers you to manage backups on your terms. Visit CometBackup.com to start your free 30-day trial today.

[00:14:11] Get $100 free credit when you sign up with the promo code MSPRADIO. Start running backups in 15 minutes or less. Comet Backup. The backup solution that MSPs trust. Thanks for listening. Today is National Employee Health and Fitness Day, National Memo Day, and National Strawberries and Cream Day. The Business of Tech is written and produced by me, Dave Sobel, under ethics guidelines posted at businessof.tech.

[00:14:41] If you've enjoyed the show, make sure you've subscribed or followed on your favorite platform. It's free and helps directly. Give us a review, too. If you want to support the show, visit patreon.com slash MSP Radio, and you'll get access to content early. Or buy our Why Do We Care merch at businessof.tech. Have a question you want answered?

[00:15:04] We take listener questions, send them in, ideally as a voice memo or video to question at MSPradio.com. I answer listener questions live on our Wednesday live show on YouTube and LinkedIn. If you've got a comment or a thought on a story, put it in the comments if you're on YouTube, or reach out on LinkedIn if you're listening to the podcast. And if you want to advertise on the show, visit MSPradio.com slash engage. Once again, thanks for listening, and I will talk to you again on our next episode.

[00:15:37] Part of the MSP Radio Network.