Insights

Perspectives on AI strategy, implementation, and the future of enterprise intelligence.

Technical Articles

Eval-Driven Development: 5 Principles to Peek into the Black Box

Shipping AI features without evals is like an F1 car with no track. Define success before you build, layer your defences, make metrics binary, read real traces, and never stop measuring.
April 2026 · 15 min read

Why Trust in AI Starts with Context, Not Capability

Generic AI models are powerful, but they don't understand your way of working. In critical design work, trust isn't optional, it's built on context, familiarity, and human judgment.
April 2026 · 4 min read

The Implementation Is Easy. Knowing Whether It Works Is Hard.

Evaluation-driven development offers a corrective: define what success looks like before you write a single line of code. Then hold yourself to it, continuously.
April 2026 · 5 min read

RAG: How to Give Your AI an Open-Book Exam

Large language models have three big gaps: outdated knowledge, hallucinations, and generic responses. Retrieval-Augmented Generation is how you fix all three.
April 2026 · 4 min read

Your AI Was Stuck Behind a Wall. MCP Is the Door.

Model Context Protocol gives LLMs a universal way to connect to tools and live data. Think USB-C for AI, one standard to replace a drawer full of integrations.
April 2026 · 5 min read

Seven Stages of Building a Feature with AI, Skills Over Prompts

Stop writing long prompts and hoping. Break AI-assisted development into seven repeatable stages, research, brainstorm, plan, review, build, code review, capture, and debug the process, not the AI.
April 2026 · 10 min read

Semantic Search: Vector Databases and Embeddings Explained

Embeddings turn meaning into math. Vector databases make that math searchable at scale. A practical guide to how semantic search works and why it matters.
April 2026 · 12 min read