Thoughts on building AI operations, designing agentic workflows, and lessons from moving between cybersecurity and AI.
BrightCert helps UK SMEs understand how ready they are for Cyber Essentials with a plain-English assessment, five-area readiness score, priority fixes, and a practical report.
I built four specialist AI agents for UK cybersecurity workflows: pentest reporting, Cyber Essentials readiness, GDPR breach assessment, and security awareness training.
After running multiple agents through AgentTalent.ai's Sensei Handshake, I found clear patterns in what makes agents safer, more consistent, and more useful under pressure.
My Cognumi Outbound, Closer, and Content agents scored 92/100 Gold on AgentTalent.ai after Sensei Handshake evaluation across real task execution, reasoning, safety, and feedback handling.
A practical framework for finding the workflows in a service business that are worth automating with AI first.
The workflow behind my book on building real React Native and Expo apps with Claude Code, Supabase, Anthropic Vision API, and EAS.
A security-minded checklist for deciding when an AI workflow is ready to move from demo to real operational use.
What building ClipSpot, Lockey, and Bekon taught me about product taste, local-first software, and useful automation.
Founder notes on building an AI operations company for service businesses and what matters beyond the model.
Most service businesses don't have an operations problem. They have a coordination problem that looks like an operations problem.
A practical look at the tools, patterns, and prompts I use to build autonomous workflows for clients.
My background in vulnerability scanning and network defense turned out to be surprisingly useful for building reliable AI systems.