{"id":168338,"date":"2025-11-30T23:28:09","date_gmt":"2025-11-30T23:28:09","guid":{"rendered":"https:\/\/www.newsbeep.com\/ie\/168338\/"},"modified":"2025-11-30T23:28:09","modified_gmt":"2025-11-30T23:28:09","slug":"could-microsoft-walk-away-with-the-corporate-ai-market-josh-bersin","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/ie\/168338\/","title":{"rendered":"Could Microsoft Walk Away With The Corporate AI Market? \u2013 JOSH BERSIN"},"content":{"rendered":"<p>I just finished diving into the announcements at <a href=\"https:\/\/www.youtube.com\/live\/TUeET4zY95c?si=MK0eNHe_1x6BxVAM\" rel=\"nofollow noopener\" target=\"_blank\">Microsoft\u2019s Ignite conference<\/a> and comparing them to the AI blitz from Google, OpenAI, and Anthropic. I have to say, at this point I think Microsoft is very likely to win a leadership position in the corporate market for AI, with a few big caveats.<\/p>\n<p>Before I explain my thinking, it\u2019s increasingly clear that the big AI vendors are now on the hunt for revenue. Having spent $2 Trillion or more on data centers, chips, and energy we can see all the big providers jockeying for position as they prepare to raise prices, driven by financial analysts ongoing complaints that AI may not be paying off.<\/p>\n<p>At the highest level it appears the market is shaking out as follows:<\/p>\n<p>Microsoft is staking out the market for corporate productivity and AI infrastructure, leveraging its massive Microsoft 365 install base and deep relationships with IT. Consider this week\u2019s announcement of <a href=\"https:\/\/news.microsoft.com\/en-hk\/2025\/11\/19\/microsoft-ignite-2025-empowering-the-frontier-firm-with-ai-agents-and-copilot\/\" rel=\"nofollow noopener\" target=\"_blank\">WorkIQ<\/a> (a system that builds inference on all the workflows, documents, and business processes in your company), the launch of <a href=\"https:\/\/www.microsoft.com\/en-us\/microsoft-agent-365\" rel=\"nofollow noopener\" target=\"_blank\">Agent 365<\/a> (a place to manage and govern agents), <a href=\"https:\/\/www.microsoft.com\/en-us\/power-platform\/topics\/app-builder\" rel=\"nofollow noopener\" target=\"_blank\">App Builder<\/a> (a visual dev environment for agents, based on Power Apps), and <a href=\"https:\/\/azure.microsoft.com\/en-us\/products\/ai-foundry\" rel=\"nofollow noopener\" target=\"_blank\">Microsoft Foundry<\/a> (previously Azure AI Studio) and its ability to manage and coordinate many models, <a href=\"https:\/\/www.anthropic.com\/news\/claude-in-microsoft-foundry\" rel=\"nofollow noopener\" target=\"_blank\">including Anthropic<\/a>.<\/p>\n<p>These are all capabilities that corporate IT departments love, because they enable companies to manage, govern, and securely deploy agents \u2013 using Microsoft embedded Copilots (more below), other agents, and <a href=\"https:\/\/ai.azure.com\/catalog\" rel=\"nofollow noopener\" target=\"_blank\">multiple models<\/a>. In a recent <a href=\"https:\/\/youtu.be\/8-boBsWcr5A\" rel=\"nofollow noopener\" target=\"_blank\">podcast by Satya Nadella<\/a>, he details Microsoft\u2019s strategy (this is well worth listening to) and explains how Microsoft\u2019s data centers are deliberately being designed for \u201cfungability,\u201d as chip sets and models change. In other words Microsoft has grown far beyond it\u2019s OpenAI beginnings: it now sees the AI market in its full and broad potential.<\/p>\n<p>For individual productivity, Microsoft has now embedded Copilot agents into Word, Excel, and Powerpoint, with insights into Outlook. For those of you who use Microsoft Office you can see how powerful this has become.<\/p>\n<p>The Excel agent, for example, understands the underlying models in Excel (it doesn\u2019t copy pages like Excel plugins for ChatGPT or Gemini) so you can ask it complex questions and it can build, discover, and fix your spreadsheets. Now imagine a similar Copilot to help you with documents, schedules, and emails. Life is going to change, that\u2019s for sure.<\/p>\n<p>On the corporate topics of AI agent observability (a new buzzword), Microsoft <a href=\"https:\/\/www.microsoft.com\/en-us\/microsoft-agent-365\" rel=\"nofollow noopener\" target=\"_blank\">Agent 365<\/a>, coupled with <a href=\"https:\/\/news.microsoft.com\/en-hk\/2025\/11\/19\/microsoft-ignite-2025-empowering-the-frontier-firm-with-ai-agents-and-copilot\/\" rel=\"nofollow noopener\" target=\"_blank\">WorkIQ<\/a>, make it pretty easy to manage all your various AI agents, and I\u2019m sure all third party providers will plug in. The company is now partnering with Anthropic to <a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/developer\/azure-mcp-server\/overview\" rel=\"nofollow noopener\" target=\"_blank\">build out MCP<\/a> (the agent-to-agent protocol we need, \u201cModel Context Protocol\u201d), and new systems called <a href=\"https:\/\/blog.fabric.microsoft.com\/en-us\/blog\/introducing-fabric-iq-the-semantic-foundation-for-enterprise-ai?ft=All\" rel=\"nofollow noopener\" target=\"_blank\">Fabric IQ<\/a> (connecting agents to databases) and <a href=\"https:\/\/techcommunity.microsoft.com\/blog\/azure-ai-foundry-blog\/foundry-iq-unlocking-ubiquitous-knowledge-for-agents\/4470812\" rel=\"nofollow noopener\" target=\"_blank\">Foundry IQ<\/a> (connecting agents to many knowledge data sources) enable IT developers to start stitching together old systems into new Superagents (<a href=\"https:\/\/joshbersin.com\/podcast\/from-assistants-to-agents-to-superagents-where-ai-is-going-next\/\" rel=\"nofollow noopener\" target=\"_blank\">listen to my podcast<\/a> for explanation of a Superagent).<\/p>\n<p>Then there\u2019s the <a href=\"https:\/\/joshbersin.com\/2025\/11\/the-josh-bersin-company-partners-with-microsoft-on-copilot-tuning-for-hr-experts\/\" rel=\"nofollow noopener\" target=\"_blank\">Copilot Fine-Tuning features<\/a>, which let you literally \u201crebuild\u201d your Copilot with embedded IP like Galileo, your company\u2019s policies and practices, and make sure that each and every employee sees a predictable and configurable experience, rather than \u201choping\u201d that ChatGPT, Gemini, or Claude behave as expected. I think this alone is enormous.<\/p>\n<p>It\u2019s an amazing array of announcements, and for me, as someone who used Windows 3.1, OS\/2, NT, and even DOS (nobody remembers, I know), it\u2019s spectacularly impressive how nimble Microsoft can be.<\/p>\n<p>As a deep observer of leadership and management, I see what a massive impact Satya Nadella has had on this company. His ability to see the future in a highly pragmatic way, coupled with his enthusiasm, energy, and open-mindedness, has propelled Microsoft into a new age.<\/p>\n<p>Remember not long ago we were laughing at Bing? Nobody\u2019s laughing at Microsoft now.<\/p>\n<p>What about Google and Gemini 3.0?<\/p>\n<p>Well right now Google is the new AI darling, effectively launching an AI platform that\u2019s every bit as useful and powerful as ChatGPT and Anthropic.<\/p>\n<p>Yes, that\u2019s an amazing feat also. Apparently Sundar Pichai did not go out and spend $20 Billion on new staff (as Meta did). Rather he shrewdly moved all the AI resources out of product groups into Gemini, building a powerhouse research and product team from within. This form of \u201c<a href=\"https:\/\/joshbersin.com\/2024\/03\/how-to-create-talent-density\/\" rel=\"nofollow noopener\" target=\"_blank\">talent density<\/a>\u201d strategy (<a href=\"https:\/\/joshbersin.com\/podcast\/understanding-talent-density-and-ditching-integrated-talent-management\/\" rel=\"nofollow noopener\" target=\"_blank\">listen to this podcast<\/a> for more) paid off, demonstrating how AI demands a very different approach to talent management.<\/p>\n<p>But Google\u2019s market is different: I think they are going after the Chief Technology Officer (not the CIO at all) and the massive software development community. Consider this: new AI vendors who automate software engineering (Microsoft GitHub, Claude Code, Cursor, OpenAI Codex, Vercel, Cognition, Replit, Lovable) are now almost a $5 billion market. Hard to believe, but Microsoft GitHub Copilot alone is Microsoft\u2019s largest AI product (around $1 billion).<\/p>\n<p>Why is this space so big? One could argue that the \u201cmanufacturing lines of the future\u201d are AI code generators. We now live in a world where 70% of GDP is driven by intellectual property, services, and deep levels of science, engineering, and process innovation. (GE\u2019s massive aviation business is not a manufacturing business, it\u2019s a business of engineering more and more efficient engines.) So why wouldn\u2019t \u201cAI coding agents,\u201d which essentially take your personal IP and embed it into scalable code, not become the lifeblood of every company?<\/p>\n<p>I used to think of software development environments (SDEs) as geeky tools for bureaucratic, over-staffed software firms. No longer. These coding copilots are tools for everyone, and even when you go to Galileo and ask for a table, you\u2019re essentially asking the AI to write code to generate the boxes and lines for you to see!<\/p>\n<p>This space, which obviously overlaps a bit with corporate (corporate IT teams are going to do most of the AI code-building), is Google\u2019s hangout. If you look at who\u2019s buying Google Cloud, you see a huge number of massive software companies, many of which will be the ERPs of the future. And given Google\u2019s more modern stack (BigQuery and other innovations which developers adore), I think this is their sweet spot.<\/p>\n<p>What about Anthropic and OpenAI?<\/p>\n<p>In the trillion dollar markets ahead, there\u2019s a lot of big thinking going on out here in Silicon Valley\/San Francisco.<\/p>\n<p>OpenAI, from what I can tell, is now one of the \u201csmaller companies\u201d in the market, yet they have a lot of tricks up their sleeve. Given that many of their new hires come from Meta, I\u2019d see them as the \u201cConsumer AI Agent\u201d that breaks out, with e-commerce, advertising, and consumer eyeballs. With 800 million users per week already, most advertisers would do anything to reach that audience with precision.<\/p>\n<p>Then there\u2019s the market for education (schools), individual creators, and all the grandparents and casual users who just want to find and learn stuff in a pleasant experience. It\u2019s almost like the Mac against PC: OpenAI is friendly, easy to use, and powerful while Microsoft is enterprise-class, scalable, and productive. We also don\u2019t know what Jony Ive will cook up, so maybe OpenAI becomes an embedded technology as well. They have lots of cool options without trying to beat down the door of CIOs.<\/p>\n<p>Anthropic, of course, is a bit different. This company has built an amazing system (we default to Claude in Galileo, although I think we\u2019ll likely switch to Gemini), and their corporate and software developer businesses are strong. Maybe Amazon eventually acquires them, giving them distribution and scale which they don\u2019t have today. I\u2019m sure they\u2019re war gaming over there, most recently standing on stage at Microsoft Ignite, pleased as punch to be part of the Azure Foundry launch.<\/p>\n<p>Oracle, Workday, ServiceNow, SAP?<\/p>\n<p>Then there are the big ERPs. They have been around for a while and they understand disruption well.<\/p>\n<p>Oracle has gotten into the data center business, a decision they may regret. Larry Ellison made a big pitch for Oracle\u2019s \u201cAI database\u201d which essentially integrates vector analysis into the SQL engine. This sounds cool but I\u2019m not sure it makes much difference, but we\u2019ll see. They are aggressively adding embedded AI into their HCM system and I would not be surprised if they build or buy an LLM themselves. They\u2019re a tenacious company and I know Larry is all over the AI play, so they\u2019ll stay relevant I promise.<\/p>\n<p>ServiceNow has a fantastic product set, staking out the space for AI-powered, application independent development tools and workflow management systems. I put ServiceNow on a plane with Microsoft as a vendor that can win over many large corporations for their pragmatic platforms, interfaces to major applications, and AI expertise. (Galileo runs inside of NowAssist and we\u2019ll be doing a roadshow with ServiceNow next year.) I trust in ServiceNow because of their culture, leadership team, and head-start on the \u201cmiddleware tools\u201d we need for AI.<\/p>\n<p>SAP is well along on their strategy and I am also very high on their approach. Not only is Joule (SAP\u2019s enterprise agent, already trained on thousands of SAP transaction flows) a groundbreaking agent for employee and business productivity, their underlying HANA (with Sybase inside) competes with Oracle, so they own the entire stack. They, like Microsoft, believe in supporting any LLM so as the LLM market consolidates (and possibly commoditizes), they\u2019re in perfect position. We all need someone to focus on our industry solutions, not just our tech stack!<\/p>\n<p>And Workday, with a new CTO and Chief Product Officer (<a href=\"https:\/\/www.workday.com\/en-us\/company\/about-workday\/leadership\/gerrit-kazmaier.html\" rel=\"nofollow noopener\" target=\"_blank\">Gerrit Kazmaier<\/a>), is becoming an AI powerhouse in a hurry. We\u2019re huge fans of Hiredscore, Paradox, and Sana \u2013 so the company now has a world-class team and amazing set of platforms to work with. In addition to \u201cagentifying\u201d Workday itself (which is well along), I believe Workday has a massive opportunity with Sana to \u201cown\u201d the learning and knowledge world of AI, even though Microsoft\u2019s WorkIQ makes noise in that direction. We\u2019re now partners with Workday on Galileo so I\u2019ll fill you in on more when we can.<\/p>\n<p>Bottom Line: AI Is Forcing Tech Vendors To Step Up Their Game<\/p>\n<p>We thought the cloud was disruptive! This wave is far more complex.<\/p>\n<p>These amazing companies now have to deal with \u201cbare metal\u201d (data centers), \u201ctoken efficiency\u201d (optimizing their cost of compute), new pricing models (no more unlimited cloud usage I\u2019m afraid), and the need to build or buy all sorts of agentic tools.<\/p>\n<p>From a strategy standpoint, each major provider cannot do everything.<\/p>\n<p>Just as Meta eventually became a consumer ad-driven business, each of these big companies has to find their \u201ccore.\u201d All I can say is that this is a very exciting time to be a business person, and I just want to celebrate Microsoft for its agility, energy, and solid execution.<\/p>\n<p>Please stay tuned for my 2026 Perspectives report, an in-depth look at corporate AI and HR for 2026 coming out in early January.<\/p>\n<p>Additional Information<\/p>\n<p><a href=\"https:\/\/joshbersin.com\/2025\/11\/gen-ai-is-going-mainstream-heres-whats-coming-next\/\" rel=\"nofollow noopener\" target=\"_blank\"> Gen AI Is Going Mainstream: Here\u2019s What\u2019s Coming Next<\/a><\/p>\n<p><a href=\"https:\/\/joshbersin.com\/podcast\/ai-economics-why-prices-are-going-up-and-the-big-shakeout-ahead\/\" rel=\"nofollow noopener\" target=\"_blank\">AI Economics: Why Prices Are Going Up And The Big Shakeout Ahead<\/a><\/p>\n<p><a href=\"https:\/\/joshbersin.com\/2025\/10\/bbc-finds-that-45-of-ai-queries-produce-erroneous-answers\/\" rel=\"nofollow noopener\" target=\"_blank\">BBC Finds That 45% of AI Queries Produce Erroneous Answers<\/a><\/p>\n<p><a href=\"https:\/\/getgalileo.ai\/?__hstc=238407455.5ecd221823b279c1e5ec71e2babbaa8a.1743986833368.1764520076390.1764523673187.167&amp;__hssc=238407455.3.1764523673187&amp;__hsfp=3416696620\" rel=\"nofollow noopener\" target=\"_blank\">Galileo: The World\u2019s Trusted Agent for Everything HR<\/a><\/p>\n<p>\u00a0<\/p>\n<p>\u00a0<\/p>\n<p>\u00a0<\/p>\n","protected":false},"excerpt":{"rendered":"I just finished diving into the announcements at Microsoft\u2019s Ignite conference and comparing them to the AI blitz&hellip;\n","protected":false},"author":2,"featured_media":168339,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[220,218,219,13634,61,60,1071,1682,80],"class_list":{"0":"post-168338","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-artificial-intelligence","8":"tag-ai","9":"tag-artificial-intelligence","10":"tag-artificialintelligence","11":"tag-google-gemini","12":"tag-ie","13":"tag-ireland","14":"tag-microsoft","15":"tag-openai","16":"tag-technology"},"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/posts\/168338","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/comments?post=168338"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/posts\/168338\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/media\/168339"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/media?parent=168338"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/categories?post=168338"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/tags?post=168338"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}