// Insights
Blog
Perspectives on legacy modernization, AI integration, and the future of enterprise technology from 35 years in the field.
AI Act Transparency Obligations: What You Must Disclose
Chatbots, deepfakes, AI-generated content — the EU AI Act requires disclosure. Article 50 creates transparency rules that apply to nearly every customer-facing AI. Here's exactly what you must do by August 2026.
Read Article →Is Your AI System High-Risk? AI Act Classification Guide
With AI Act obligations taking effect August 2, 2026, every enterprise must know if their systems are affected. The 8 domains of Annex III, the 5-question test, and concrete compliance obligations.
Read Article →Personal, De-identified, and Anonymized Data Under Quebec's Law 25
The distinction most organizations miss: de-identified data is still personal information under Law 25. It remains fully regulated. Anonymized data is not. Here's what that means for your AI systems, with concrete examples.
Read Article →COBOL PARK: Japan Just Created a Company to Save COBOL — Here's Why That Matters Everywhere
FPT and SCSK launched COBOL PARK to address the COBOL developer shortage. With the average mainframe developer over 55 and mass retirements accelerating, this isn't just Japan's problem—it's a global infrastructure emergency that affects every enterprise running legacy systems.
Read Article →Why 85% of Enterprise AI Projects Fail — And How Legacy Systems Are the Hidden Culprit
The real reason most enterprise AI never reaches production isn't the model — it's legacy infrastructure never designed for real-time, API-driven workloads. Here's what enterprises get wrong at the integration layer, and the bridge approach that actually works.
Read Article →The CISA C/C++ Deadline Has Passed: What Enterprises Should Actually Do Now
The January 2026 memory safety deadline wasn't about dropping C++. It was about publishing a roadmap. Here's what CISA actually required, why panic-migrating to Rust is the wrong move, and a practical 6-step playbook for hardening your legacy code.
Read Article →Why 85% of Enterprise AI Projects Fail (And How to Succeed)
The problem isn't the AI—it's integration, compliance, and unrealistic expectations. Here's what the 15% that succeed do differently, and how to make AI actually work in production.
Read Article →C and C++ Still Run the World's Critical Infrastructure
60% of critical infrastructure runs on C and C++—from banking systems to medical devices to industrial control. Here's why these languages remain irreplaceable and what it means for enterprises.
Read Article →COBOL in 2026: Why It Still Runs the World
95% of ATM transactions, 80% of in-person transactions, and $3 trillion in daily commerce still run on COBOL. Here's why this "dead" language refuses to die—and what it means for enterprise modernization.
Read Article →Stay Updated
Get Legacy & AI Insights
No spam. Just practical insights on enterprise technology.
Subscribe