Post-Quantum Cryptography: What to Do Before the Deadline Feels Real
Nobody can tell you when a quantum computer will break RSA, and you don't need to know. Data you encrypt today can be stored and read later, and the migration takes years. Here's where a sensible programme starts.
Continue Reading
Small Models, Big Wins: When On-Device AI Beats the Cloud
Not every AI feature needs a frontier model and a trip across the network. Small language models running on laptops, phones and in the browser can now handle a surprising amount of real work, and they change the maths on privacy, speed and cost.
Continue Reading
Observability Without the Enterprise Price Tag
When production breaks, the first question is "what happened?", and small teams often can't answer it. OpenTelemetry has made proper tracing, metrics and logs affordable, and you can get the basics running in a week.
Continue Reading
Your Cloud Bill Is Mostly Leftovers
Cloud overspend rarely comes from one bad decision. It builds up from forgotten environments, oversized servers, surprise data-transfer charges and logs nobody reads. Where to look first, roughly in order of payoff.
Continue Reading
Your Site Loads Fast. So Why Does It Feel Slow?
Your speed scores can look great while real users tap a button and wait. Interaction to Next Paint (INP) measures that lag. It's a Core Web Vital, and it's where a lot of modern sites fall short.
Continue Reading
Build for No Signal: The Case for Offline-First Apps
Field staff work in basements, paddocks, plant rooms and on rural roads. If the app shows a spinner every time the signal drops, they'll stop using it. Offline-first design treats the network as a bonus, not a requirement.
Continue Reading
You Probably Don't Need Another Database
Search engine, queue, document store, vector database: modern architectures add a new service for every job. For most business apps, one well-run Postgres can do all of it, with far less to operate, secure and pay for.
Continue Reading
Passkeys Are Ready. Is Your Customer Portal?
Passwords are the weakest part of most customer portals and the biggest source of support tickets. Passkeys are phishing-resistant, quicker to use and supported on almost every device. The tech is ready. The rollout is where the work is.
Continue Reading
Your AI Pilot Worked. Here's Why It Still Won't Ship.
Most enterprise AI pilots succeed. Most never make it to production. The gap usually isn't technical. It's a handful of unglamorous decisions nobody made at the start.
Continue Reading
Prompt Injection Is Not a Theoretical Problem Anymore
Any AI agent that reads untrusted content can be hijacked by it. The fix isn't a cleverer prompt. It's the same boring security architecture you'd use for any system that handles hostile input.
Continue Reading
Rolling Out AI Coding Tools Without Wrecking Your Codebase
GitHub Copilot, Claude Code and Cursor can double a team's output, of good code and bad code alike. Whether you get the upside or a pile of review debt depends on everything around the tool.
Continue Reading
Where AI Cost Actually Goes in 2026 (and What to Do About It)
Token prices keep falling, yet enterprise AI bills keep climbing. Most of the money goes on architecture choices that never appear on a pricing page, and most of it can be won back.
Continue Reading
MCP: The Quiet Standard That's Finally Making AI Agents Useful
Model Context Protocol has become the standard way to plug AI agents into business tools and data. What MCP is, why it matters, and how to start using it in your own environment.
Continue Reading
Long Context Is Here: What Claude 4 Changed for Enterprise Work
Million-token context windows have changed what's possible in enterprise AI. What Claude 4 really changed, where long context beats retrieval, and where it still doesn't.
Continue Reading
How to Build Custom AI Agents with Microsoft Copilot Studio
A practical guide to building business AI agents in Microsoft Copilot Studio: how to scope them, what to connect them to, and how to keep them useful after launch.
Continue Reading
Enterprise AI Readiness: A Practical Assessment Framework
Before you spend on AI, check whether your organisation is ready for it. The framework I use with clients to assess data, infrastructure, skills, governance and use cases.
Continue Reading
Claude vs Gemini vs GPT: Choosing the Right AI for Your Enterprise
A hands-on comparison of Anthropic Claude, Google Gemini and OpenAI GPT for enterprise work: where each one is strongest, and the factors that usually decide the choice.
Continue Reading
From Legacy to AI-Powered: A Migration Playbook
How to bring AI into legacy business systems a step at a time, without a full rewrite. Practical patterns for adding new capability around the applications you already run.
Continue Reading