Asurgent AB
Högskolan på Gotland
Greater Stockholm Metropolitan Area
2 tn följare
Fler än 500 kontakter
2 tn följare
Fler än 500 kontakter
Info
Certifierad lösningsspecialist med huvudfokus inom automation och effektivisering…
Artiklar av Robin
Aktivitet
2 tn följare
-
Robin Iwarzon Erlandsson har omdelat dettaRobin Iwarzon Erlandsson har omdelat dettaAI-Ready Architecture Series (Part 3): AI Gateway, MCP, and Reducing the Attack Surface of AI Agents In Part 1, I talked about reducing the attack surface of AI agents using VNet integration and private connectivity. In Part 2, I covered why API Management becomes the control plane for AI-ready APIs. Now the next questions are: 👉 “What about the AI Gateway in API Management?” 👉 “And where does MCP actually fit?” The important thing to understand is this: The AI Gateway in Azure API Management isn’t a separate product. It’s APIM doing what it already does well, but applied to AI workloads: • Agents calling APIs • Agents invoking tools • Models interacting with enterprise systems It gives you a central place to: • Enforce authentication and authorisation • Apply policy and rate limiting • Control and monitor traffic • Reduce the exposure of backend systems This pattern is stable and production ready today: Copilot / Agent → APIM → API / backend system That gives you: • A single controlled entry point • No direct exposure of internal services • Full audit and traceability of agent actions • A significantly reduced attack surface Microsoft has also introduced AI Gateway integration directly within Foundry, backed by API Management. While this is useful for getting started, enterprise architectures will often benefit from bringing their own APIM instance so AI workloads inherit: • Existing network boundaries • Private endpoints and VNet integration • Shared governance and observability • Existing security controls MCP is where things get interesting. APIM can now: • Expose existing MCP servers • Surface APIs as MCP-style tools • Route MCP traffic through the gateway This moves the model from: • APIs as endpoints To: • Tools as governed capabilities But the important point is: 👉 MCP does not replace APIM 👉 It builds on top of the same governed architecture AI agents don’t become valuable just because they can call APIs. They become valuable when they can securely interact with enterprise systems at scale without increasing your attack surface. That starts with governance, not tooling. #Azure #APIManagement #PowerPlatform #CopilotStudio #ZeroTrust #SecOps #AgenticAI #AzureAI #CloudArchitecture #ANSGroup
-
Robin Iwarzon Erlandsson har omdelat dettaRobin Iwarzon Erlandsson har omdelat dettaExcited to share this practical guidance on designing workloads on Microsoft Fabric through the lens of the Well-Architected Framework. No matter the type of workload, there are always important decisions to make around capacity, data design, governance, collaboration, and operations. Check out the recommendations for building production-ready Fabric workloads while making the most of the platform. 🔖 https://lnkd.in/g4VRDMAH Huge thanks to Slava Trofimov, Omer A. Nicholas Hurt, Katie Murray, Ian S., Clayton Siemens for bringing your product expertise and real-world experience to help shape this guidance. Martin Ekuan, Cassandra Oxford, Josée M., Reut Maimon #MicrosoftFabric #WellArchitectedFramework #CloudArchitecture #DataArchitecture #Analytics #DataEngineering
-
Robin Iwarzon Erlandsson har omdelat dettaRobin Iwarzon Erlandsson har omdelat dettaGitHub just declared war on vibe coding. With their own repo. 100k stars. MIT. 206 contributors. 146 releases. It's called Spec Kit. The idea => specifications become executable. You describe what to build, the agent implements from the spec. No more prompting and hoping. Seven commands run the entire workflow: → /speckit[.]constitution - governing principles and development guidelines → /speckit[.]specify - what to build, what and why, no tech stack required → /speckit[.]clarify - finds the gaps before planning starts → /speckit[.]plan - technical implementation plan with your stack → /speckit[.]analyze - consistency check across spec, plan, and tasks before any code runs → /speckit[.]tasks - actionable task list → /speckit[.]implement - executes everything Tasks become GitHub Issues via /speckit[.]taskstoissues. Quality checklists via /speckit[.]checklist. 30+ AI coding agents supported. Claude Code, Codex, Copilot, Cursor, and more. Extensions add new commands. Presets override templates, terminology, and standards without touching tooling. Here's the GitHub Repo: https://lnkd.in/g5vGi4EP
-
Robin Iwarzon Erlandsson har omdelat dettaRobin Iwarzon Erlandsson har omdelat dettaMicrosoft just open-sourced Azure Skills — a capability layer for coding agents on Azure. 25 skills across deploy, diagnostics, cost, RBAC, AI, AKS — paired with MCP tooling for live Azure context and actions. This is the infrastructure play: give every AI agent first-class Azure fluency out of the box. The agents that win won't just write code. They'll deploy, monitor, and secure it too. 🔗 https://lnkd.in/gguBCeb8
-
Robin Iwarzon Erlandsson har delat dettaGood morning Emma. A few months ago I decided to pick up the cursed subject of On/Offboarding. Everyone needs it, but everyone is also different and usually some of it is already covered in an internal system that wants to be master. We will see where this attempt lands but one new addition that I previously haven't tried out is the User facing portal. The what? You know the portal that actually collects vital information for your new team mate and presents it in one place. Ah that one, we can't prioritize it. Previously this part of the solution has been super hard to motivate, as it primarily does not return any value on the rest of the build. It is simply super nice to have, but not vital or core. Now though with agentic engineering we can put it in there without risking taking to much effort/time/investment from the rest of the solution. And the best part, it will be self-sustained as we can now utilize agentic workflows to both pull, present and refine key data. Let's keep building #onboarding #offboarding #agenticengineering #solution
-
Robin Iwarzon Erlandsson har omdelat dettaRobin Iwarzon Erlandsson har omdelat dettaMicrosoft just dropped an Agentic Transformation Patterns Playbook ✨ Here's what's inside 👇 1️⃣ Six transformation patterns, from Employee AI Enablement and Workplace & IT Services through to Core Business Processes and customer-facing agents 2️⃣ A detailed 5×5 maturity model to see where you stand today 3️⃣ The Agentic Center of Excellence model that closes those gaps and helps scale agents safely 4️⃣ Your 90-day plan to move from pilots to production, starting Monday The bit I keep thinking about: ➡️ Your weakest capability becomes your ceiling. Find the one gap that breaks scale for your pattern. Fix that first. 🔗 Link in the comments to save a copy 👇 Which of the six patterns is your team running today, and what's the one capability gap holding you back? #AI #Agents #MicrosoftCopilot #FutureOfWork Manuela Phil
-
Robin Iwarzon Erlandsson har delat dettaCreating illustrations with AI image generation and refining its details can be a time sink-hole. We have previously addressed how AI tools can very, very rapidly give you a 60% maybe 70% quality output but then aggressively exponentially scale in time consumption to improve even single digits. This is the hype trap. Basically you build up your source material inside maybe a project, then add context to align expectations and finally when you are brave or curious enough you let it run. Normally you output back that is at first glance, surprisingly impressive. You get hyped and travel way ahead of your actual progress. Next you dig in and start analysing your one shot output and here is where it gets complicated. Overall it looks great, but small or sometimes bigger details are off. You now need to refine and refining with AI is messy. During the past year I've explored a side project together with my oldest daughter. We are building and producing a children's book that will be published as an open source component on both Github and at home https://lnkd.in/diFnPBk9 It will be free to use, play or just enjoy together. As a kid my uncle came up with imaginare bed time stories whenever he visited. I enjoyed it a lot and have kept the tradition within the family. This started off as one of those. I then got curious what technology could assist with. I've always enjoyed writing and story telling on a novice level but can't draw at all. Looking forward to sharing more about our wip progress and the book itself, but until then here is some real world example of how you very, very quickly can get results but then spend A LOT of time refining to barely inch close to a finalized product.
-
Robin Iwarzon Erlandsson har delat dettaAre you a Microsoft oriented customer and want to build business Agents but struggle to get started? I hope this can assist. It has been a couple month now since I last addressed a bigger topic but lately I've been seeing a lot of different views on where to get started if you want to build and deploy business agents. I've done a fair share of it and hope this sum up can be of assistance or at least save you some time and effort. Lately I've been busy building and one of my focus areas have been to refresh https://iwarzon.com , converting it from portfolio to a Lab workspace where we can share and engage. Anyway, feel free to give it a try and please reach out if you find anything that looks odd or have any thoughts on improvement or subjects you want me to address. It has been a while since I was in the web-game. Looking forward to sharing more. Let's keep pushing together. #agents #microsoft #lab #building #engineering
-
Robin Iwarzon Erlandsson har omdelat dettaRobin Iwarzon Erlandsson har omdelat dettaAI Gateway Architecture on Microsoft Azure — Simplified Blueprint 🚀 Most teams jump into AI without a clear architecture. That’s where things break. Here’s a practical blueprint to build a scalable, secure AI gateway on Microsoft Azure: 🔹 Identity & Security First Use Azure AD (Entra), Key Vault & Managed Identity to keep everything secure by design. 🔹 Central Gateway with APIM Azure API Management acts as the control layer—routing, throttling, and governing all AI requests. 🔹 Event-Driven Processing Azure Event Hub enables async + batch workloads for high-scale scenarios. 🔹 Seamless AI Integration Connect to Azure OpenAI PTUs or LLM deployments via secure network layers. 🔹 Observability Built-In Azure Monitor + Logs + Metrics → full visibility into usage, performance, and failures. 🔹 Actionable Insights Dashboards, alerts, and automated actions ensure reliability at scale. 💡 Why this matters? Because AI systems without governance = chaos. This architecture ensures control, scalability, and production readiness. If you're building AI systems on Azure, this is the layer most people miss. 📌 I break down AI concepts like this every week - plus how to actually BUILD. Free. No jargon. Just clarity. 👉 Subscribe to The AI Revenue Playbook: https://avsl.beehiiv.com/ Follow Aiswarya Venkitesh for more AI insights.
Robin Iwarzon Erlandsson har kommenterat ett inlägg
1 v
https://iwarzon.com/blog/creating-illustrations-with-ai-image-generation
Robin Iwarzon Erlandsson har kommenterat ett inlägg
1 v
Link to article: https://iwarzon.com/blog/select-agentic-platform-to-build-agents
Robin Iwarzon Erlandsson har kommenterat ett inlägg
2 v
Grattis Boulder till en mycket fin rekrytering. Fortsätt innovera och leverera Christian Björnegren! 🚀
-
Robin Iwarzon Erlandsson har reagerat på dettaRobin Iwarzon Erlandsson har reagerat på dettaMany organizations already have copilots, analytics, and data in place, but how to connect these into everyday business operations? For me, one of the most interesting shifts right now is the rise of the 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗹𝗮𝘆𝗲𝗿: combining data, AI, governance, and business processes into something that actually supports operational work instead of isolated demos or use cases. This also changes how solutions are built. 𝗧𝗵𝗲 𝗳𝘂𝘁𝘂𝗿𝗲 𝗶𝘀 𝗶𝗻𝗰𝗿𝗲𝗮𝘀𝗶𝗻𝗴𝗹𝘆 𝗵𝘆𝗯𝗿𝗶𝗱 — 𝗰𝗼𝗺𝗯𝗶𝗻𝗶𝗻𝗴 Fabric, Data Agents, Copilot Studio, Azure services, integrations, and governance models together instead of treating them as separate worlds. We’ll discuss these themes together with Heidi Korhonen in our upcoming Forward Forever webinar: 📅 𝗙𝗿𝗼𝗺 𝗗𝗮𝘁𝗮 𝘁𝗼 𝗗𝗲𝗰𝗶𝘀𝗶𝗼𝗻𝘀: 𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝘁𝗵𝗲 𝗘𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗔𝗜 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗟𝗮𝘆𝗲𝗿 10.6.2026 at 10:00–10:45 𝗧𝗼𝗽𝗶𝗰𝘀 𝗶𝗻𝗰𝗹𝘂𝗱𝗲: • Fabric, Data Agents, and Microsoft IQ services • Hybrid architecture and connected business processes • Governance as part of enterprise AI architecture • Moving from insights toward operational actions 👉 Register here: https://lnkd.in/gNrBFUVc #MicrosoftFabric #AIAgents #EnterpriseAI #CopilotStudio #MicrosoftCopilot #AIAgents #Microsoft365 #CopilotAgents #Governance #DigitalTransformation #ForwardForever #PowerPlatform #SharePoint #AIGovernance #Security
-
Robin Iwarzon Erlandsson har reagerat på dettaRobin Iwarzon Erlandsson har reagerat på dettaNu är det dags att ändra status till #opentowork! Tack Annika Berthold för förtroendet att få vara med att säkra och utveckla varuförsörjningen inom Region Stockholm genom rollen som interim CIO på MediCarrier. Det har varit både spännande och väldigt roligt att få leda IT genom en förändringsresa mot ett säkrare, mer strukturerat och framför allt ett robustare IT. Nu lämnar jag med varm hand över kaptenshatten till Per Hansson som fortsätter arbetet som permanent CIO. Behöver ni en interim CIO efter sommaren som kan driva förändring, stärka IT-leveransen och skapa struktur i komplexa miljöer är ni varmt välkomna att höra av er. #jagab #cio #itchef #interimcio #interimitchef #interim
-
Robin Iwarzon Erlandsson har reagerat på dettaRobin Iwarzon Erlandsson har reagerat på dettaSom en del av sin pågående digitala transformation har Diaverum inlett en omfattande modernisering av sin centrala IT‑plattform genom att flytta verksamhetskritiska system till Microsoft #Azure. Initiativet – Project Zenith – stärker organisationens skalbarhet, säkerhet och driftseffektivitet och skapar en enhetlig, global plattform för framtida innovation. I nära samarbete med sitt ägarbolag och Microsoft har Diaverum valt Asurgent som partner för att etablera målmiljön i Azure och leda den stegvisa migreringen från den tidigare servermiljön till molnet. Läs hela caset här: https://lnkd.in/dZVfY4CaDiaverum kundcase - Modernisering av global infrastruktur i Azure - AsurgentDiaverum kundcase - Modernisering av global infrastruktur i Azure - Asurgent
-
Robin Iwarzon Erlandsson har reagerat på dettaEfter 8 år på Bengt Dahlgren har det blivit dags för mig att sätta kurs mot nya utmaningar. ✈️ Jag är otroligt tacksam för allt förtroende jag fått genom åren, alla spännande projekt och alla fantastiska kollegor jag lärt känna längs vägen. Något jag tar med mig är tankesättet ”Från människa till samhälle” - en bra vägvisare, inte minst i dessa tider, både personligen och som ett långsiktigt perspektiv för affärer. Det finns otroligt många skickliga kollegor inom Bengt Dahlgren, och det engagemang och den passion ni visar för det ni gör är verkligen inspirerande. Tack för allt jag fått lära mig tillsammans med er. Från och med onsdag denna vecka fortsätter jag min digitala resa som Cloud Consultant på Asurgent AB, där jag kommer hjälpa deras kunder att förbättra processer och utveckla verksamheter med fokus på automation och AI. Tack för ett väldigt varmt välkomnande och en grym pre-boarding av Melina Söderberg, Ofelia Bordier med flera. 🥳
-
Robin Iwarzon Erlandsson har reagerat på dettaRobin Iwarzon Erlandsson har reagerat på dettaAI-Ready Architecture Series (Part 3): AI Gateway, MCP, and Reducing the Attack Surface of AI Agents In Part 1, I talked about reducing the attack surface of AI agents using VNet integration and private connectivity. In Part 2, I covered why API Management becomes the control plane for AI-ready APIs. Now the next questions are: 👉 “What about the AI Gateway in API Management?” 👉 “And where does MCP actually fit?” The important thing to understand is this: The AI Gateway in Azure API Management isn’t a separate product. It’s APIM doing what it already does well, but applied to AI workloads: • Agents calling APIs • Agents invoking tools • Models interacting with enterprise systems It gives you a central place to: • Enforce authentication and authorisation • Apply policy and rate limiting • Control and monitor traffic • Reduce the exposure of backend systems This pattern is stable and production ready today: Copilot / Agent → APIM → API / backend system That gives you: • A single controlled entry point • No direct exposure of internal services • Full audit and traceability of agent actions • A significantly reduced attack surface Microsoft has also introduced AI Gateway integration directly within Foundry, backed by API Management. While this is useful for getting started, enterprise architectures will often benefit from bringing their own APIM instance so AI workloads inherit: • Existing network boundaries • Private endpoints and VNet integration • Shared governance and observability • Existing security controls MCP is where things get interesting. APIM can now: • Expose existing MCP servers • Surface APIs as MCP-style tools • Route MCP traffic through the gateway This moves the model from: • APIs as endpoints To: • Tools as governed capabilities But the important point is: 👉 MCP does not replace APIM 👉 It builds on top of the same governed architecture AI agents don’t become valuable just because they can call APIs. They become valuable when they can securely interact with enterprise systems at scale without increasing your attack surface. That starts with governance, not tooling. #Azure #APIManagement #PowerPlatform #CopilotStudio #ZeroTrust #SecOps #AgenticAI #AzureAI #CloudArchitecture #ANSGroup
-
Robin Iwarzon Erlandsson har reagerat på dettaRobin Iwarzon Erlandsson har reagerat på dettaExcited to share this practical guidance on designing workloads on Microsoft Fabric through the lens of the Well-Architected Framework. No matter the type of workload, there are always important decisions to make around capacity, data design, governance, collaboration, and operations. Check out the recommendations for building production-ready Fabric workloads while making the most of the platform. 🔖 https://lnkd.in/g4VRDMAH Huge thanks to Slava Trofimov, Omer A. Nicholas Hurt, Katie Murray, Ian S., Clayton Siemens for bringing your product expertise and real-world experience to help shape this guidance. Martin Ekuan, Cassandra Oxford, Josée M., Reut Maimon #MicrosoftFabric #WellArchitectedFramework #CloudArchitecture #DataArchitecture #Analytics #DataEngineering
Erfarenhet och utbildning
Se Robins fullständiga erfarenhet
Se titel, anställningstid med mera.
eller
Licenser och certifikat
-
-
-
MS-101 Microsoft 365 Mobility and Security
Microsoft
UtfärdatLegitimerings-id https://www.credly.com/earner/earned/badge/4d6644a3-8379-4ab4-bfbf-33c32a377637 -
-
-
-
-
-
Kurser
-
ITIL Foundation
-
-
Projekt och Ledarskap
-
Utmärkelser och priser
-
Microsoft MS-500 Security Administrator Associate
Microsoft
Earning Microsoft 365 Security Administrator Associate certification validates the skills and knowledge to proactively secure Microsoft 365 enterprise and hybrid environments, implement and manage security and compliance solutions, respond to threats, and enforce data governance.
Språk
-
Engelska
Professionell yrkeskunskap
-
Svenska
Modersmåls- eller tvåspråkig nivå
-
Danska
Grundläggande kunskaper
-
Norska
Begränsad yrkeskunskap
Mottagna rekommendationer
1 person har rekommenderat Robin
Gå med nu för att se