Published Articles
AI Workflows vs AI Agents: Where to Set the Dial
The workflow-vs-agent choice is not binary. It is a dial - and most enterprises set it too far right. Cost, reliability, and EU AI Act oversight all favour the middle for most enterprise work.
5 Workflow Patterns That Solve 95% of Enterprise AI Problems
Prompt chains, routing, parallelisation, orchestrator-subagent and evaluator loops. The gap between one LLM call and a full agent is where 95% of production value lives - and most teams jump straight past it.
When You Genuinely Need an Agent: A Production Readiness Checklist
The 12 questions to answer before granting autonomous decision-making to any system.
The Agentic CLV, Part 1: From Metric to Live Signal
CLV sits on slides and rarely shapes what happens on a Tuesday afternoon. Agentic AI is changing that - moving lifetime value from a lagging report to a live signal that autonomous systems consume and act on.
The Modelling Stack - Real-Time CLV in Production
The three levels of CLV sophistication, what changes when predictions must serve in milliseconds, and the MLOps realities of always-on scoring.
The Operating Model - Structuring Around Customer Trajectories
Feeding CLV back into ad platforms, redesigning retention around the days-0-to-90 cliff, and the organisational shifts required.
EU AI Act: A Practical Implementation Guide for CTOs
One obligation is already active and most organisations missed it. A non-generic guide to AI inventory, classification, the Betriebsrat dimension, and what the August 2026 deadline means for your planning calendar.
Context Engineering: The Missing Piece in Enterprise Agentic AI
The 6-layer framework, context rot, and the maturity curve that separates AI demos from production deployments.
Building Production-Ready Enterprise AI Agents
Four architecture decisions separate agents that work in demos from agents that work reliably at scale. Memory layers, context compression, human oversight design, and multi-agent orchestration.