AI & Automation
Agents, LLMs, and enterprise AI — a space for practitioners.
Coming Soon
60 articles on SAP, AI, architecture, and enterprise tech — each with an AI reading companion that has read the piece. Ask questions, challenge claims, go deeper while you read.
The Interactive Blog
Publishing has been a monologue for thirty years. This piece explains the interactive blog you're about to use.
5 min readDiscuss →Most pilots die between demo and production — organizational, predictable, and fixable.
6 min readDiscuss →Boundaries, audit trails, and humans in the architecture.
7 min readDiscuss →Why clean core went from marketing slide to hard prerequisite.
6 min readDiscuss →What self-hosting actually buys you — and what it does not.
6 min readDiscuss →The dirty secrets of SAP project planning no consultant will tell you upfront
8 min readDiscuss →Spoiler: it wasn't LangChain
7 min readDiscuss →A field guide to surviving SAP's biggest architectural shift
9 min readDiscuss →GPU bills, hallucinations, and the dream of free AI inference
8 min readDiscuss →Orchestration patterns, failure modes, and lessons from 6 months in production
10 min readDiscuss →Because vendor marketing is not your friend
7 min readDiscuss →The skills that matter, the ones that don't, and the meetings you should skip
6 min readDiscuss →A decision framework for people who actually need results
9 min readDiscuss →Why enterprise UX is stuck and what it takes to fix it
6 min readDiscuss →The techniques that work, the ones that only work in demos, and the gap between them
8 min readDiscuss →Vector databases sound magical until you actually use one
11 min readDiscuss →The hidden costs, hidden benefits, and what to actually negotiate
7 min readDiscuss →Framing technical depth for executive conversations
6 min readDiscuss →Monitoring, alerting, and the automation mindset that changed how I work
7 min readDiscuss →A skeptic's take on the AI agent hype cycle
8 min readDiscuss →A practitioner's read on what actually shipped that matters, what's still noise, and the one structural bet worth making for the next 18 …
9 min readDiscuss →The integration layer every AI stack will need — explained without the hype.
8 min readDiscuss →Test-time compute is real — and expensive. A practitioner's routing guide.
9 min readDiscuss →Local inference only works when you run the box like ops, not like a gaming PC.
8 min readDiscuss →Yesterday's guess becomes today's fact unless you design memory like a database, not a diary.
9 min readDiscuss →Demos show one perfect frame. Production gets 400-page scans and blurry phone photos.
8 min readDiscuss →When everyone's base model is 'good enough,' data, routing, and ops become the product.
9 min readDiscuss →Bad chunks make genius models look stupid. Fix the scissors first.
8 min readDiscuss →Practical adversarial testing for teams who can't hire a purple-team consultancy.
7 min readDiscuss →Big context fixes retrieval problems you should solve with design — and creates new ones.
9 min readDiscuss →Escalation beats defaulting to the biggest model — same lesson as batch vs. interactive.
8 min readDiscuss →From zero to one real tool surface in an afternoon — no framework tourism.
10 min readDiscuss →Governance that blocks demos is theater. Governance that blocks production incidents is insurance.
7 min readDiscuss →Demos are montages. Production is a single narrow workflow that must not fail.
7 min readDiscuss →Joule isn't your enemy — unplanned duplication of agents, data, and governance is.
9 min readDiscuss →Swap the chat model all day — if retrieval misses, the answer was dead on arrival.
8 min readDiscuss →If developers mute the bot, you don't have review — you have spam.
8 min readDiscuss →Users don't want a conversation with expense policy — they want a button that works.
9 min readDiscuss →Edge AI isn't miniature cloud — it's different jobs with different success metrics.
8 min readDiscuss →If they can't answer in a shared doc, the answer is no.
7 min readDiscuss →The agent is a user with amnesia and superpowers — design permissions like you mean it.
9 min readDiscuss →SAP's ABAP MCP Server is generally available after Sapphire 2026. A practitioner's guide to wiring agentic ABAP development into your S/4 program without blowing your AI Units budget.
6 min readDiscuss →SAP ABAP Development Tools for VS Code is generally available in 2026. Scope limits, MCP integration, and when your team still needs Eclipse ADT.
6 min readDiscuss →Claude Code vs Cursor isn't either/or in 2026. How a technical architect pairs terminal agents and IDE flow for SAP, infra, and side projects.
6 min readDiscuss →Sub-agents, hooks, and skills are Claude Code's secret orchestration layer. Patterns for multi-step automation without runaway token bills.
6 min readDiscuss →Cloudflare Agents SDK now supports MCP clients and Durable Objects on the free tier. Build and deploy a production AI agent at the edge without a Kubernetes bill.
6 min readDiscuss →Stateful AI agents on Cloudflare need Durable Objects done right. Session isolation, SQL persistence, and memory patterns that don't contaminate tomorrow's answers.
6 min readDiscuss →SAP AI Units went commercial in 2026. How to forecast Joule and ABAP AI consumption, control costs, and know when to use your own stack instead.
6 min readDiscuss →SAP bet big on Claude and MCP for Autonomous Enterprise. What shipped, what's roadmap, and how architects should evaluate without sitting through another keynote.
6 min readDiscuss →Understanding the strategic difference between ayraix.com's AI Hub (curated intelligence) and Community Blog (practitioner insights) — and why both are essential.
7 min readDiscuss →Where agents help in brownfield migration, where humans must sign off, and rollback triggers no model should own.
8 min readDiscuss →A comprehensive guide to securing Model Context Protocol connections in SAP environments, covering authentication, authorization, data protection, and monitoring essentials.
7 min readDiscuss →LiteLLM transforms your homelab into a flexible AI gateway with fallback chains, load balancing, and cost control across local and remote LLMs.
7 min readDiscuss →Visual workflows, retries, credentials, and cron hygiene for agent pipelines — without inventing a custom scheduler per job.
8 min readDiscuss →Replace traditional VPNs with NetBird's zero-trust mesh network for secure, seamless access to your homelab AI services from anywhere.
7 min readDiscuss →Strategies for sharing a single RTX GPU between Ollama LLM inference and ComfyUI Stable Diffusion on the same homelab machine.
7 min readDiscuss →How RISE with SAP customers can leverage AI-powered migration assistants to reduce risk, effort, and disruption during their move to the cloud.
7 min readDiscuss →How SAP's 2021+ Cloud Private Edition releases created the foundation for today's AI capabilities, establishing the technical prerequisites that made large-scale AI deployment possible.
7 min readDiscuss →How the technical architect role evolved from diagram creation to AI agent governance, oversight, and ethical guidance in the agentic era.
7 min readDiscuss →Transform your TrueNAS SCALE NAS into a full AI operations platform running LLMs, Stable Diffusion, vector databases, and orchestration workflows.
7 min readDiscuss →Coffee Corner
Phase 2: rooms where readers discuss with each other, not just with the AI.
Agents, LLMs, and enterprise AI — a space for practitioners.
Coming SoonLeadership, transformation, and building durable companies.
Coming SoonTech regulation, AI governance, and industry policy.
Coming SoonTimely topics — analyzed and discussed together.
Coming SoonArticle requests, feedback, Coffee Corner ideas — [email protected]