Markus Bosch

How to seamlessly integrate your #AI agents, #api and sources? Just do what Microsoft is doing and get aboard of Model Context Protocol (#MCP): https://lnkd.in/eC7K9ntP This is a protocol first presented by Anthropic in November 2024: https://lnkd.in/e3ns7gm8

7

This is cool. I’ve been saying for a while that the problem with putting MQTT brokers in the middle of architectures is that very few systems leverage only “current state”. Many need access to history and this is a great solution to that problem.

13

1 Kommentar

𝐀 𝐂𝐚𝐫𝐞𝐞𝐫 𝐌𝐢𝐬𝐭𝐚𝐤𝐞 𝐈 𝐒𝐭𝐢𝐥𝐥 𝐒𝐞𝐞 𝐏𝐞𝐨𝐩𝐥𝐞 𝐌𝐚𝐤𝐢𝐧𝐠 𝐢𝐧 𝐓𝐞𝐜𝐡 𝐒𝐭𝐚𝐲𝐢𝐧𝐠 𝐜𝐨𝐧𝐟𝐢𝐧𝐞𝐝 𝐭𝐨 𝐚 𝐬𝐢𝐧𝐠𝐥𝐞 𝐭𝐞𝐜𝐡𝐧𝐨𝐥𝐨𝐠𝐲 𝐨𝐫 𝐫𝐨𝐥𝐞 𝐟𝐨𝐫 𝐭𝐨𝐨 𝐥𝐨𝐧𝐠. Even today, many IT professionals remain deeply rooted in one platform, one language, or one toolset, waiting years to “master” it before exploring something new. That approach might have worked a decade ago. But in today’s fast-paced world of AI, cloud-native architectures, DevSecOps, and automation-first environments, being too specialized can quietly turn into being outdated. 𝐓𝐡𝐞 𝐫𝐞𝐚𝐥𝐢𝐭𝐲 𝐢𝐬: ✅ Businesses expect tech professionals to wear multiple hats ✅ Security challenges span across domains ✅ Collaboration between infra, app, and security teams is the norm ✅ Soft skills—like communication, stakeholder management, and adaptability—are no longer optional; they’re essential 𝐖𝐡𝐚𝐭’𝐬 𝐚 𝐛𝐞𝐭𝐭𝐞𝐫 𝐚𝐩𝐩𝐫𝐨𝐚𝐜𝐡? 🔹 Broaden your technical perspective—understand how your stack fits into the bigger architecture 🔹 Learn the fundamentals of IT infrastructure, cloud, APIs, business apps, or automation, even if it’s outside your current role 🔹 Practice translating technical complexity into business-relevant language 🔹 Don’t wait to be told to reskill, stay ahead of where tech is heading 𝐈𝐧 𝐈𝐓 𝐚𝐧𝐝 𝐈𝐧𝐟𝐨𝐒𝐞𝐜, 𝐲𝐨𝐮𝐫 𝐯𝐚𝐥𝐮𝐞 𝐢𝐬𝐧’𝐭 𝐣𝐮𝐬𝐭 𝐢𝐧 𝐰𝐡𝐚𝐭 𝐲𝐨𝐮 𝐤𝐧𝐨𝐰, 𝐢𝐭’𝐬 𝐢𝐧 𝐡𝐨𝐰 𝐟𝐚𝐬𝐭 𝐲𝐨𝐮 𝐜𝐚𝐧 𝐞𝐯𝐨𝐥𝐯𝐞. #Cybersecurity #ITCareers #CareerGrowth #TechSkills #SoftSkillsMatter #TechLeadership #ContinuousLearning #Adaptability #ITLife #ITInfrastructure #CoachAPM

10

1 Kommentar

A new thinktank paper on nine technology trends is extremely to the point: ...most of these shifts hit organizations now or within the next five years and will force “Peter”, the standard employee, to fundamentally adapt how he works – from executing tasks to supervising AI, handling exceptions and owning outcomes. At the same time, organizations must rapidly adjust structures, skills and governance so that people and technology evolve together rather than drifting apart – and the best part is that the paper explains this in a very easy-to-understand way while outlining a seamless roadmap for how to get started.

2

The AI hype is over. And that’s a good thing. Today at the Microsoft AI Tour in Munich it was obvious: AI has become a commodity. No more magic, just another layer of infrastructure we’re expected to run and govern.Most companies I see already have a zoo of tools in place: GPT here, Copilot there, plus Claude, Gemini and a couple of “strategic pilots” in the business. Everyone talks about harmonisation, but in reality we’re adding new tools faster than we clean up the old ones.Waiting for one dominant platform to solve this for us is, in my view, a trap: There won’t be a one‑size‑fits‑all AI stack.The edge won’t come from “the next model”, but from architecture, governance and measurable impact. Curious how others see this in 2026: Are you trying to standardise on a single AI platform – or do you embrace a multi‑vendor stack and focus on getting the governance layer right?

11

Exciting updates from #OneDrive in #Microsoft365! Check out the latest features enhancing productivity and organization: #CopilotAgents: Build custom AI assistants within OneDrive for seamless document and project management. #AISummaries: Instant summaries and FAQs from file content in OneDrive and SharePoint, boosting efficiency. #FileComparison: Easily compare Word, PowerPoint, and PDF files to pinpoint differences quickly. #SmartViews: Intelligently categorize files and photos for swift access and better organization. #MobileMemories: Explore the new OneDrive mobile app with immersive albums and natural language photo search. #DesignerIntegration: Edit images effortlessly in OneDrive with background removal and text addition. #AdminTools: Advanced reporting, site-policy comparison, and AI-driven content governance. #ArchiveMagic: File-level archiving in SharePoint libraries for streamlined content lifecycle management. Stay ahead with these innovative tools and features from OneDrive! 🚀 https://lnkd.in/e38vuEnA #Productivity #AIIntegration #OneDrive #Copilot

12

🚀 Agentic AI is rising fast—and so is the urgency to act with clarity and cohesion. As organizations explore the transformative potential of Agentic AI, one thing is clear: speed alone isn’t enough. To truly unlock value, we need both a clear strategy and agile execution. Strategy defines your North Star. Agile execution helps you take iterative steps, learn, and adapt along the way. But here’s the catch: no strategy will succeed in silos. Too often, we see fragmented efforts—teams working in isolation, duplicating work, or missing critical perspectives. In this new era, collaboration isn’t optional—it’s foundational. Bringing the right stakeholders together early, aligning on goals, and building with shared purpose is what will separate the leaders from the laggards. Let’s stop treating AI as a side project. Let’s treat it as a shared mission. #AgenticAI #DigitalTransformation #Collaboration #Strategy #AILeadership #EnterpriseAI #Innovation #FutureOfWork

30

1 Kommentar

#AIGrowthMindset 🔍 Is AI a Threat or an Opportunity? Let’s be honest—it’s not even a fair question anymore. AI is an inevitability. The only choice we have is how we respond to it. Over the last three decades, technology has radically transformed industries. From mainframes to the internet, from robotic automation to IoT—we’ve seen wave after wave of disruption. And now, AI is here, not as a tool—but as a tectonic shift. 🚀 Look around you: #Tech giants Microsoft moved from license selling to enabling outcomes for industries through cloud services. #Travel agents became travel apps—convenience at your fingertips (MakeMyTrip, Booking. com). #Pharma R&D timelines have been cut in half, improving drug affordability so accessibility at scale. #Retail From malls to mobile—Amazon, Flipkart, Blinkit, Best Buy changed the game. The lesson is simple: Technology adoption isn't optional anymore. You can’t run tomorrow’s business using yesterday’s playbook. 💡 AI is not just another trend—it’s a strategic inflection point. And like every revolution, it demands a bold mindset shift—from everyone in the C-Suite. To stay relevant and lead transformation, business leader must adopt an "AI Growth Mindset". And digital leader must act as co-pilots to the CEO—helping envision, shape, and execute the future business model. 🧭 Ask yourself: What does your business look like 5–8 years from now? 1. Can you reimagine your positioning against competitors? 2. Can AI help to design new products, services, and business models? 3. Can you reimagine your future customers, their value proposition? 4. What will be profit model/engine (subscriptions, pay-per-use) look like? 5. What would be your operating model (touchless, intelligent plants)? 6. Who all could be your future partners? Are you building the right ecosystem of partnerships to scale? 📌 The question isn’t “Is AI a threat or opportunity?” The question is: Are you ready to lead in the age of intelligence? Let’s stop watching the disruption and start owning it. #ChiefDigitalOfficer #AILeadership #DigitalTransformation #FutureOfBusiness #GrowthMindset #Innovation #AIAdoption #Leadership #CxOInsights

15

🧠 What is Agentic AI — and why it actually matters Have you heard of Agentic AI? It’s not just another smart chatbot or content-generator. It’s a game changer. 🔹 Generative AI creates content when you ask — text, images, code. 🔹 Agentic AI goes a step further: it sets goals, plans the steps, acts on its own and adapts — often across many tools and tasks. Think of the difference like this: Generative AI is a “content machine”, while Agentic AI is more like a “digital coworker” — someone who can take initiative and actually get things done. That shift? It might completely change how we work, automate tasks — maybe even how companies operate. #AgenticAI #GenAI #AI #FutureOfWork

7

⏰ Pharma Prescribed Podcast episode alert with Nives Rombini, Co-Founder of tech start-up Navis Bio, a venture dedicated to harnessing data-driven insights for biomedical innovation.   🎓 Her professional journey began with Swisscom, where she honed her skills in data engineering and AI applications, before transitioning to the pharmaceutical sector.   Fluent in German, Italian, English, and French, Rombini embodies a cosmopolitan perspective that enriches her scientific collaborations.   🩺 With a foundation in both laboratory research and computational analysis, she is recognized for her ability to connect disciplines, close evidence gaps, and inspire innovation at the intersection of biology and data science. 🎙️ This is a must listen to for anyone who wants to hear and see energy from someone at the forefront of innovation. 👇 Links in the comments below. #pharmaprescribedpodcast #aimedtech #aipharma #navisbio

31

3 Kommentare

Microsoft Copilot rated “good” in independent test. SRF’s Kassensturz asked the University of Applied Sciences Northwestern Switzerland (FHNW) to evaluate ten AI chatbots across 331 real‑world tasks. Microsoft Copilot received an overall “good” rating — with strong performance in health questions, Swiss law and creative tasks. AI adoption in Switzerland continues to accelerate. Our latest AI Diffusion Report shows that about one in three working‑age people already use AI tools at work — well above the global average. The results highlight where Copilot delivers real, everyday value for people in Switzerland — professionally and privately. At Microsoft Switzerland, our goal is clear: build AI that empowers people, respects privacy and strengthens our digital future. Reviews like this help us deliver on that promise for our customers and partners. 👉 Read the full study: https://msft.it/6045tU5GS #MSFTemployee

6

The future of AI is autonomous agents-intelligent systems that go beyond static models to reason, plan, and act independently. This comprehensive guide breaks down the architecture of AI agents into three core components: the reasoning model (the brain), actionable tools (the hands), and the orchestration layer (the nervous system). It presents a taxonomy from simple connected problem-solvers to self-evolving multi-agent systems that learn and adapt over time. The article highlights key design principles for building production-grade agents, including robust security models, governance at scale, and the importance of operational rigor through Agent Ops-a new discipline for managing AI unpredictability. Advanced examples like Google Co-Scientist and AlphaEvolve showcase the transformative potential of agents in research and problem-solving. For developers, architects, and product leaders, this work provides the foundational blueprint to build AI agents that are not just automations but collaborative, adaptive members of teams driving real-world impact. #AI #AutonomousAgents #GenerativeAI #AgentArchitecture #AIInnovation #AgentOps

23

#Microsoft #Copilot could do so much better. Currently the experience with Copilot is the little AI than can't. 9 out of 10 things I try to ask it just flop. Team's Copilot, much better. The Microsoft team has to consider how to empower people to perform tasks within Excel (at all), but I would have said better. The good news is, that is a bunch of work and a lot of jobs for people to apply AI to the Excel tool, to make it easier for those people who use Excel. It's just right now, Copilot resembles today's best measure of how far everyone else has moved forward. I wonder what CoPilot for PowerBI will be like. Is it even worth it to explore? --- Watching Copilot try to come back with an answer is like watching grass grow.

3

𝐄𝐱𝐜𝐢𝐭𝐢𝐧𝐠 𝐧𝐞𝐰𝐬 𝐟𝐫𝐨𝐦 #𝐌𝐢𝐜𝐫𝐨𝐬𝐨𝐟𝐭𝐁𝐮𝐢𝐥𝐝 𝟐𝟎𝟐𝟓! 🚀 Microsoft is pushing the boundaries of AI and developer tools with groundbreaking updates unveiled in #MSBuild. Here are some highlights: 🌟 𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝟯𝟲𝟱 𝗖𝗼𝗽𝗶𝗹𝗼𝘁 𝗧𝘂𝗻𝗶𝗻𝗴: A new low-code capability in Copilot Studio lets organizations customize AI models with their own data and workflows—no data science team required basic tasks! 🌐 𝗔𝘇𝘂𝗿𝗲 𝗔𝗜 𝗙𝗼𝘂𝗻𝗱𝗿𝘆: A unified platform for building AI apps and agents, now supporting over 1,900 models, including xAI’s Grok 3. 🤖 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗖𝗼𝗽𝗶𝗹𝗼𝘁 & 𝗠𝟯𝟲𝟱 𝗔𝗴𝗲𝗻𝘁𝘀 𝗧𝗼𝗼𝗹𝗸𝗶𝘁: Enhanced AI agents for Microsoft 365, with new SDKs and APIs to streamline development and integration. 📝 𝗘𝗱𝗶𝘁 𝗼𝗻 𝗪𝗶𝗻𝗱𝗼𝘄𝘀: An open-source, lightweight command-line text editor (<250KB) for developers, rivaling tools like vim. 🛠️ 𝗪𝗶𝗻𝗱𝗼𝘄𝘀 𝗦𝘂𝗯𝘀𝘆𝘀𝘁𝗲𝗺 𝗳𝗼𝗿 𝗟𝗶𝗻𝘂𝘅 (𝗪𝗦𝗟) 𝗢𝗽𝗲𝗻-𝗦𝗼𝘂𝗿𝗰𝗲𝗱: WSL’s source code is now on GitHub, inviting community contributions for enhanced features. 🆓 𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗦𝘁𝗼𝗿𝗲 𝗙𝗲𝗲 𝗪𝗮𝗶𝘃𝗲𝗿: Individual developers can now publish apps on the Microsoft Store for free, a first for global digital storefronts 🔗 𝗢𝗻-𝗗𝗲𝘃𝗶𝗰𝗲 𝗔𝗜 𝗳𝗼𝗿 𝗘𝗱𝗴𝗲: New APIs bring Microsoft’s Phi-4-mini model to web apps, enabling features like text generation and summarization. Check out the full details in the Microsoft Build 2025 Book of News: https://lnkd.in/ekHrH9Za What’s your favorite update? #MicrosoftBuild #AI #Developers #Innovation #MSBuild #TechNews #Microsoft365

12

𝗜 𝗵𝗮𝘃𝗲 𝘄𝗶𝘁𝗻𝗲𝘀𝘀𝗲𝗱 𝗵𝗼𝘄 𝗱𝗶𝗴𝗶𝘁𝗮𝗹 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝗲𝘃𝗼𝗹𝘃𝗲𝗱 𝗳𝗿𝗼𝗺 𝗽𝗵𝘆𝘀𝗶𝗰𝗮𝗹 𝗿𝗲𝗮𝗹𝗶𝘁𝘆 𝘁𝗼 𝗼𝗿𝗯𝗶𝘁𝗮𝗹 𝗽𝗼𝘀𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆. Early in my career, infrastructure was tangible. Servers you could hear. Heat you could feel. Decisions that directly impacted business continuity. Then virtualization reshaped the paradigm. Infrastructure became less about hardware and more about architecture. Flexibility started to matter as much as stability. Resilience was never theoretical. I worked on real disaster recovery capabilities that required subsea connectivity to keep operations running across regions. In those moments, infrastructure was not just an IT topic. It was a business survival priority. Later came the cloud wave. Many organizations believed they could fully eliminate their own data centers. Cloud did transform the operating model. But infrastructure did not disappear. It evolved. At the same time, the industry began exploring ideas that once sounded like science fiction. Microsoft tested underwater data centers to improve cooling efficiency and reliability. https://lnkd.in/eag9uUZJ Today the conversation is shifting again. The explosive growth of artificial intelligence is forcing leaders to rethink the scale, location and energy model of computing. Some concepts now explore moving compute beyond Earth. Proposals linked to SpaceX suggest constellations of solar powered satellites could support distributed workloads in orbit. https://lnkd.in/eaFxuQ2i Space offers near continuous solar energy and natural thermal conditions that could enable new levels of computing scalability. https://lnkd.in/e9fwFK6G 𝗔𝗳𝘁𝗲𝗿 𝗺𝗮𝗻𝘆 𝘆𝗲𝗮𝗿𝘀 𝗹𝗲𝗮𝗱𝗶𝗻𝗴 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝘁𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 𝗶𝗻𝗶𝘁𝗶𝗮𝘁𝗶𝘃𝗲𝘀, 𝗼𝗻𝗲 𝗰𝗼𝗻𝘃𝗶𝗰𝘁𝗶𝗼𝗻 𝗶𝘀 𝗰𝗹𝗲𝗮𝗿. Infrastructure is no longer a background capability. It is becoming one of the defining strategic choices organizations make. It shapes innovation speed. It shapes resilience. It shapes competitiveness in an AI driven economy. 𝗔𝗿𝗲 𝘄𝗲 𝗱𝗲𝘀𝗶𝗴𝗻𝗶𝗻𝗴 𝗽𝗹𝗮𝘁𝗳𝗼𝗿𝗺𝘀 𝗳𝗼𝗿 𝘁𝗵𝗲 𝗲𝗺𝗲𝗿𝗴𝗶𝗻𝗴 𝘄𝗼𝗿𝗹𝗱, 𝗼𝗿 𝘀𝘁𝗶𝗹𝗹 𝗼𝗽𝘁𝗶𝗺𝗶𝘇𝗶𝗻𝗴 𝘁𝗵𝗲 𝗼𝗻𝗲 𝘄𝗲 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗸𝗻𝗼𝘄? #TechnologyLeadership #ArtificialIntelligence #DataCenters #Cloud #DigitalTransformation #Innovation

23

💬 𝗔 𝘀𝗶𝗺𝗽𝗹𝗲 𝘁𝗼𝗼𝗹 𝘁𝗼 𝗲𝘅𝗽𝗹𝗼𝗿𝗲 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁 𝗔𝗜 𝗺𝗼𝗱𝗲𝗹𝘀 𝗮𝗻𝗱 𝗼𝗻𝗹𝘆 𝗽𝗮𝘆 𝗳𝗼𝗿 𝘄𝗵𝗮𝘁 𝘆𝗼𝘂 𝘂𝘀𝗲 - 𝘁𝗿𝘆 𝗪𝗶𝘁𝘀𝘆 I recently came across and tried out a free desktop tool (available for Windows, MacOS und Linux) called 𝗪𝗶𝘁𝘀𝘆, and I think it’s an interesting option if you're working with large language models and want more flexibility. In essence, it looks very similar to other interfaces like ChatGPT. It is a clean interface where you can connect 𝗺𝗼𝘀𝘁 𝗼𝗳 𝘁𝗵𝗲 𝗺𝗮𝗷𝗼𝗿 𝗔𝗜 𝗺𝗼𝗱𝗲𝗹𝘀 𝘃𝗶𝗮 your own 𝗔𝗣𝗜 𝗸𝗲𝘆 - all in one place. That includes things like: •  OpenAI’s GPT-4o •  Google’s Gemini 2.5 Pro •  And others like Claude, Mistral, etc. You simply enter your API keys from OpenAI, Google etc. and pay based on your actual (token) usage. So, no extra subscription or license fees, which makes it handy if you don’t want to spend $20/month on every single subscription. 👀 𝗜 𝘀𝗲𝗲 𝘁𝘄𝗼 𝗺𝗮𝗶𝗻 𝘂𝘀𝗲 𝗰𝗮𝘀𝗲𝘀: 1️⃣ You want to test and compare different LLMs 2️⃣ You’re using just one provider (like ChatGPT), but prefer a pay-per-use setup instead of a flat monthly fee (which can be a lot cheaper depending on your usage). 𝗪𝗵𝗮𝘁’𝘀 𝗮𝗹𝘀𝗼 𝗰𝗼𝗼𝗹 𝗶𝘀 𝘁𝗵𝗮𝘁 𝗪𝗶𝘁𝘀𝘆 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝘀𝘂𝗽𝗽𝗼𝗿𝘁𝘀 𝘀𝗼𝗺𝗲 𝗻𝗶𝗰𝗲 𝗳𝗲𝗮𝘁𝘂𝗿𝗲𝘀 𝗼𝘂𝘁 𝗼𝗳 𝘁𝗵𝗲 𝗯𝗼𝘅: •  RAG (Retrieval-Augmented Generation) to build your own knowledge base. •  Shortcuts to use the text in any application as input for your model. •  Specialized agents (or in ChatGPT language "GPTs") like an accountant or analyst. •  Image and video generation if your model supports it. •  Audio transcription – live during the chat or by uploading a file. •  And there’s even long-term memory support (although in my test, it didn’t carry over between sessions – but it’s there in the settings). If you're more privacy-conscious, you can even 𝗿𝘂𝗻 𝗶𝘁 𝘄𝗶𝘁𝗵 𝘆𝗼𝘂𝗿 𝗼𝘄𝗻 𝘀𝗲𝗹𝗳-𝗵𝗼𝘀𝘁𝗲𝗱 𝗟𝗟𝗠 𝘃𝗶𝗮 𝗢𝗹𝗹𝗮𝗺𝗮, and the data stays in your environment. Overall, I think it's a great tool to explore different models, compare outputs, and try out more advanced features. #GenAI #AItools #Witsy #APIs #LLMs #RAG #Productivity #BYOK

8

4 Kommentare