# Blog · Decision capability and governance How companies become capable of deciding again. Articles on Business Data Strategy, Governance, the EU AI Act and operational friction. Dargestellte Fassung: https://simo-online.com/en/blog SIMO Blog · Decision knowledge for executive management ## Executive Management SIMO Blog Decision knowledge for Articles on Business Data Strategy, Governance and decision-making capability. Evidenced, current, applicable. Articles Categories Timeframe 2025–26 Search articles … All Newest first Oldest first Blog articles ## No articles found Try different search terms or reset the filters. Clear filters Back to the overview Category Author Published on Reading time Tags Source: LinkedIn profile (opens in a new tab) CTO and Co-Founder, SIMO GmbH Mathematician (Diplom) with more than 20 years of responsibility for data and transaction architectures at major German banks. Advises regulated institutions on Business Data Strategy and Governance with an eye on GDPR and the EU AI Act. Co-Founder, SIMO GmbH Co-Founder of SIMO GmbH. Supports mandates in the Mittelstand and in large enterprises, focused on corporate strategy, Change Management and alignment at board level. SIMO GmbH Editorial Team The editorial team is made up of senior consultants of SIMO GmbH, led by Andreas O. Schwan. Focus areas: Business Data Strategy, Business Data Management, Governance, the EU AI Act and the GDPR. Read by topic Eight topics group the articles. Each one carries its own framing: why it matters for steering a company. Blog Articles in this topic All articles ## Tools decide nothing. Architecture decides. Platforms, interfaces, models, and the question of what gets switched off in return. Articles on platforms, interfaces and models, read from the vantage point that counts in the boardroom. Every tool decision is an architecture decision with a long tail. A connector, a model, an orchestration layer: what starts as a pilot today sits in operations, in the budget and in the audit trail two years from now. The articles in this topic explain how RAG, the Model Context Protocol, LLM orchestration and the wiring into ERP and CRM actually work, and where the running costs appear that no vendor quote mentions. SIMO does not pick tools. We settle the questions that come first: which business question the tool is meant to answer, who owns the resulting figure, and which system disappears in exchange. A company that adds a tool every year and retires none is buying complexity on installments. Fewer systems, not more: less a cost measure than the precondition for decisions that can still be traced. Taking tools seriously means planning the shutdown as well. AI Tools & Technology · Platforms, interfaces and models How tool decisions become architecture decisions: RAG, the Model Context Protocol, LLM orchestration and the wiring into ERP and CRM, and what belongs on the shutdown list. ## Data is a board matter. Not a tooling matter. Data quality, availability, ownership: the ground every decision stands on. Articles on data preparation, modeling, real-time analysis and the question of who owns a number. Data quality is a symptom, not a cause. Treating it as a project means cleaning tables while leaving the condition that produced them untouched: nobody in the business owns the number that matters. The articles here run from data preparation through statistical modeling to real-time analytics, and at every stop they show that the bottleneck is rarely the technology. A data strategy answers three questions before any system is chosen: which decision is supposed to improve, which number carries it, and who in the company owns that number by name. Only then does architecture become a sensible question. That order is the difference between a data strategy and a data project, and it belongs on the executive agenda, not in a department. From the number to the name behind it: that is the work. Information & Data Management · Data quality, modeling, ownership Why data quality is a symptom and who owns a number: articles on data preparation, statistical modeling and real-time analytics. ## Strategy starts with the business. Not with the model. Why programs fail on leadership rather than on technology. Articles on target states, economics and funding, and on the reasons programs stall. One figure repeats across almost every article in this topic: most AI programs never reach production. The cause is rarely the model. It is a missing target state, numbers without an owner, and data nobody has examined for the purpose it is suddenly asked to serve. A strategy is not a shopping list. It states which decision is meant to improve, by when, what that is worth, and how you would recognize the opposite. These articles supply the material for it: return calculations, funding programs, data quality as a foundation, leadership as the bottleneck. What they do not supply is a shortcut. There is none here. Where SIMO works, and when another firm is the better address. AI Strategy · Target state, economics and leadership Why programs fail on leadership rather than technology: articles on target states, return calculations, funding and data quality as the foundation. ## Do not automate a detour. Shorten it. Procurement, accounting, sales, supply chain: the routines that eat the day. Articles on routines in procurement, accounting, finance, sales and the supply chain. Automation is the most visible form of relief and the most treacherous. An automated detour is still a detour. It merely runs faster, and it is harder to change afterward because an interface now depends on it. The articles here work through concrete routines: e-invoicing and bookkeeping, sourcing and supplier scoring, inventory, sales. So the same question comes first every time: why does this step exist? If it can go, the automation is saved. If it stays, it deserves to be described and owned before anything automates it. Transformation is meant to remove complexity. A program that leaves behind more moving parts than it found has missed its purpose. How to orchestrate routines without breeding the next tangle. AI Automation · Procurement, accounting, sales and supply chain Why an automated detour is still a detour: articles on e-invoicing, sourcing, inventory, sales and no-code platforms. ## Regulation is not a project. It is an operating condition. EU AI Act, GDPR, shadow AI: obligations with deadlines and fines. Articles on the EU AI Act, the GDPR, AI governance and Germany's implementing act. The EU AI Act, the GDPR and the German implementing act do not ask for one push. They ask for evidence that holds at any moment: which systems are in use, in which risk class, on which data, under whose responsibility. The articles here sort the obligations by deadline and say what each one means in practice, including shadow AI, the usage nobody registered. A company does not become auditable through a policy. It becomes auditable through traces: decisions that can be followed, owners with names, documented data lineage. Those traces are produced in daily operations or not at all. Which is why regulation belongs in the architecture rather than in a binder. What the EU AI Act actually requires, ordered by deadline. Compliance & Regulation · EU AI Act, GDPR and AI governance Obligations with deadlines instead of policies in a binder: articles on the EU AI Act, the GDPR, shadow AI and auditable governance. ## Competence is not optional. It is required. Article 4 of the EU AI Act makes training mandatory. Here is what that means. Articles on learning paths, prompting practice, training formats and the literacy duty under Article 4. Since Article 4 of the EU AI Act, training is no longer a nice to have: anyone deploying AI systems has to make sure the people working with them understand them. The articles here compare the formats that deliver on this (course, certificate, workshop, coaching) and describe how to tell when a session only ticks a box. Enablement is at once the cheapest part of a transformation and the first to be cut. A company that rolls out tools and trains nobody shifts the load onto a few individuals and becomes dependent on them. Knowledge held in a single head is not an asset; it is an exposure. Understand and build autonomous AI agents, hands-on. AI Education & Skills · The literacy duty under Article 4 of the EU AI Act What Article 4 of the EU AI Act requires in practice: articles on learning paths, prompting practice and a comparison of training formats. ## Industries differ. The questions do not. Trades, retail, construction, healthcare: same questions, different conditions. Articles from the trades, retail, construction, healthcare and professional services. Every industry brings its own conditions: the trade shop schedules crews, the retailer steers inventory, the clinic answers for patient data, construction runs against dates. The articles here work through those differences with concrete examples: figures, entry points, and the places where things reliably jam. What does not differ is the order of business: which decision should improve, which number carries it, who owns it. An industry example does not replace that clarification; it only makes it tangible. Borrow a template from another industry and you borrow its assumptions with it. How SIMO works in your industry. Industry Solutions · Trades, retail, construction and healthcare Same questions, different conditions: articles from the trades, retail, construction, healthcare and professional services. ## New is not an argument. Evidence is. Digital twins, agents, quantum machines: evidence over enthusiasm. Articles on digital twins, autonomous agents, quantum computing and green AI. Each of these technologies arrives with a market forecast and a date. The articles here quote both and then ask the question behind them: what is demonstrably running today, and what is an announcement? Digital twins run in manufacturing, autonomous agents are reaching applications, quantum machines sit in test centers. The distance between those statements is the actual content. For an executive team the useful question is not what will be possible in 2030, but what a decision taken today will be worth in two years. So we judge emerging technology by evidence: systems in operation, figures that were checked, owners with names. Everything else is a bet someone else pays for. Evidence instead of announcements: what we have actually delivered. Future Technologies · Digital twins, AI agents, quantum computing What demonstrably runs today and what is an announcement: articles on digital twins, autonomous AI agents, quantum computing and green AI.