Paul van der Linden

Two weeks ago, I spoke with a new leader in Enterprise Architecture who shared an interesting perspective. His EA team (8 people) had been covering both Enterprise Architecture and Business Architecture responsibilities. On paper, that sounds efficient. In reality, he felt it wasn’t working. He felt that capability models that leaned too heavily toward IT enablement and gap analyses that missed business operations. Communication he felt was often reactive to situations without any elements of control. With a maximum potential headcount of 10 across both EA and Business Architecture, his plan is clear: 👉 First, assess the performance of his existing EA team. 👉 Then, create a dedicated Business Architecture function The infographic we created below outlines the difference between a Business Architect and an Enterprise Architect doing a dual role (EA+BA). It’s a reminder that while there is overlap, the focus and outputs are often very different. #Architecture #BusinessArchitecture #EnterpriseArchitecture #EA #Strategy #Business #Communication #Leadership

202

54 commentaren

If you are a M365 customer in Germany, Austria, Switzerland or anywhere in Europe, this update is for you. Please read it. Microsoft has just introduced #Anthropic models in #Copilot, and there are some important details you need to know before you enable them 🚨 Anthropic’s Claude Opus 4.1 is now available alongside OpenAI, but only in the Researcher agent and in Copilot Studio additional to Claude Sonnet 4 today. Great optionality, limited day to day impact for now. Microsoft announced in a side note, that this model availability of #Anthropic will come to M365 Copilot Chat experiences too! My clear CTA today: Switch it off 🔴 (or today: do not opt-in!) Why it matters Switching to Anthropic routes your prompts and responses OUTSIDE YOUR Microsoft’s managed environment. That means several enterprise protections you expect with Copilot do not apply. To be clear: no Microsoft Product Terms and DPA no Data residency and EU Data Boundary commitments no Microsoft audit and compliance controls and SLAs no Microsoft Customer Copyright Commitment My take for production tenants: Do not enable this yet. Block Anthropic at tenant level until protections reach parity. If you want to explore, do it only in a non production sandbox with non sensitive data and tight admin controls. By contrast, Microsoft‑hosted Microsoft 365 Copilot continues to operate under Microsoft’s privacy, security, and compliance commitments (including EU Data Boundary support). That’s the baseline many organisations chose Copilot for! Even in Copilot Studio, Anthropic models are treated as external and are not for production deployments How to block in the admin center Admin center > Copilot > Settings > Data access > AI providers for other large language models > Anthropic > Block provider My Bottom line Model choice is the right direction. Today, it is a promising preview for risk-aware companies, not a green light for regulated or risk averse organisations, especially in the EU. How are you approaching this change: pilot now or wait for protection parity? Raphael Köllner

175

39 commentaren

🔄 New in Copilot Studio: Smarter Session Outcome Detection in Generative Orchestration 🤖 Until recently, makers had to explicitly trigger the 'End of Conversation' topic in Copilot Studio to avoid sessions being marked as abandoned. That meant if a user went silent, even after the agent provided a helpful response, the session could be incorrectly flagged as abandoned. We’ve changed that. 💡 With the new update in Generative Orchestration mode, Copilot Studio now automatically evaluates the session after 15 minutes of inactivity: If there are no active plans (i.e., nothing is waiting for user input), the session is flagged as ✅ resolved. If an agent asked a question and the user doesn’t reply, it’s still marked as ❌ abandoned. In short: if your agent did its job and there’s nothing left to do, we now recognize that as a success—even without explicit confirmation. 💬✅ https://lnkd.in/d6qFP6VX #CopilotStudio #analytics #ConversationalAI Eran Manor Gary Pretty

211

21 commentaren

🚀 𝐓𝐡𝐞 𝐉𝐨𝐮𝐫𝐧𝐞𝐲 𝐭𝐨 𝐄𝐧𝐭𝐞𝐫𝐩𝐫𝐢𝐬𝐞 𝐀𝐫𝐜𝐡𝐢𝐭𝐞𝐜𝐭𝐮𝐫𝐞 𝐌𝐚𝐭𝐮𝐫𝐢𝐭𝐲 Enterprise Architecture (EA) is not a one-time project, but a strategic journey that moves from chaos to full optimization. This model outlines the five key levels of maturity, showing how organizations can evolve their systems and processes to enable continuous innovation. → Level 1: Initial At this early stage, EA is either nonexistent or highly unstructured. Systems and processes operate independently, leading to inefficiencies, redundancy, and misalignment across the organization. → Level 2: Repeatable Some processes and frameworks start to take shape, but they remain siloed and inconsistent. While teams may reuse certain methodologies, there's no overarching governance or integration between systems. → Level 3: Defined At this point, EA becomes more formalized, with defined standards, integrated systems, and alignment with business objectives. Processes are documented and followed across departments. → Level 4: Managed EA is now actively monitored and optimized using data-driven decision-making. Organizations establish KPIs to measure efficiency, performance, and alignment with business goals. → Level 5: Optimized At the highest level of maturity, EA is fully embedded into business strategy. It enables continuous innovation and provides a competitive advantage. 📢 At YALLO Group, through our Talent in a Box solution, we connect top professionals in Enterprise Architecture with organizations to help them move up the maturity model and achieve their business goals. #EnterpriseArchitecture #EA #DigitalTransformation #MaturityModel #yallo.co

48