-
OpenClaw vs. Workflow Engines: Autonomous Agents vs. Deterministic Flows
OpenClaw and workflow engines like n8n, Make, or Scrydon are often mentioned together, but they solve fundamentally different problems. This post explores the architectural divide between autonomous AI agents and deterministic flow-based platforms, and where the two are converging.
-
The Future of Agentic AI is Flow-Based
After nearly two years of working intensively with Agentic AI—first as Sr CSA Manager Data & AI at Microsoft, working with strategic enterprise customers, and now as CEO of Scrydon—I’ve come to a clear conclusion: the future of Agentic AI is flow-based.
The initial promise of fully autonomous agents—AI systems that could independently reason, plan, and execute complex tasks—has given way to a more nuanced reality. As McKinsey notes in their analysis “One year of Agentic AI: Six lessons from the people doing the work”:
It’s not about the agent; it’s about the workflow.
This insight resonates deeply with what I’ve observed across dozens of enterprise deployments. Let me explain why flow-based agentic AI represents the path forward.
-
AI Revolution Unleashed using AI Foundry at AI Tour Brussels '25
-
Interactive Home Assistant notifications: Charge your car with a Tap
Home Assistant is an incredibly powerful platform for automating your home, and one of its standout features is the ability to create interactive notifications. Imagine this: it’s 21:30, you’re home, and your electric car isn’t charging yet. Home Assistant can send a notification to your phone asking if you’d like to start charging at a specific time—and with a single tap, you can confirm it. In this blog post, I’ll walk you through setting up a notification with an actionable response to control your car charger (and potentially other tasks) using Home Assistant automations.
-
Small Language Models (SLM) at Tide '25