About · az365.ai
Alex Pechenizkiy
Microsoft AI for business applications, applied surgically. Where it adds value, not where it demos. Two decades shipping on the Microsoft business apps stack: D365, Power Platform, Dataverse, CRM, XRM, custom apps. 250+ projects across 6 countries. ML since 2005. The math, the delivery, and the discipline to know which AI belongs where.
Why I write about Azure and AI
Microsoft AI for business applications, applied surgically. Where it adds value, not where it demos. Two decades shipping on the Microsoft business apps stack: D365, Power Platform, Dataverse, CRM, XRM, custom apps. 250+ projects across 6 countries. ML since 2005. The math, the delivery, and the discipline to know which AI belongs where.
Multi-AI by design. Claude when Claude is better, OpenAI when OpenAI is better, Foundry when integration wins. Most Azure bloggers are Azure-only loyalists. I pick the model that fits the workload, not the badge wall.
That combination is what this site is. Honest evaluation of Microsoft's AI surface, grounded in what actually ships.
What I bring to the table
Microsoft business apps
Two decades shipping on D365, Power Platform, Dataverse, CRM, XRM, and custom operational apps. 250+ projects across 6 countries. The stack, end to end.
AI & Machine Learning
ML since 2005. Now evaluating and integrating Azure AI services into production. I know what the model can deliver, what the launch demo says, and the gap between.
Surgical AI
AI applied where it adds value, not where it demos. Multi-AI by design: Claude when Claude is better, OpenAI when OpenAI is better, Foundry when integration wins.
The Full Stack
C#/.NET, TypeScript, React, Python, Power Platform, Dataverse. Architecture whiteboard to working code. That matters when evaluating AI tools.
Why az365.ai exists
Microsoft is betting everything on AI. Azure AI Foundry, Copilot across every product, AI Builder in Power Platform, autonomous agents. The pace is relentless. The documentation is comprehensive but sanitized. You find the "how" but rarely the "should you."
Where are the licensing gotchas? The features that demo well but break at scale? The architectural decisions that lock you in? The cost math on AI consumption that can blow past your budget in a week?
That is the gap this site fills. I write from the perspective of someone who has to make these things work in production, not sell them at Ignite. If something is not ready, I say so. If there is a cheaper way, I show you the numbers.
What I write about
- Azure AI services. AI Foundry, OpenAI, Document Intelligence, Cognitive Services. What works, what does not, what it actually costs.
- Copilot & AI agents. Microsoft 365 Copilot, Copilot Studio, autonomous agents. Real-world evaluation, not launch-day hype.
- Cloud architecture. Azure patterns for AI workloads, governance, security, cost optimization.
- AI adoption in the enterprise. How organizations actually integrate AI into existing Microsoft stacks (Power Platform, D365, M365).
- The math. Licensing, consumption costs, ROI analysis. The numbers Microsoft does not put in the brochure.
- Build vs buy vs wait. When to adopt, when to build custom, when something is just not ready yet.
Credentials
Personal Microsoft certifications. The architecture and AI-relevant set is shown below; the full list lives on Credly .
Architecture & AI
Development & Platform
MSc in System Engineering. ML research background since 2005. Founder, Richlode Solutions (2010-2018): Microsoft Dynamics CRM and SugarCRM consulting practice.
Get in touch
Building on Azure? Evaluating AI services? Trying to figure out if Copilot is worth the seat price? Always happy to talk shop.