What I’m exploring now
A connected research programme, run hands-on in a private lab, asking one question: how will technology organisations, software delivery and executive technology leadership change when intelligent agents become active participants in the enterprise?
Agentic software factories
Hekton is a private software factory where fleets of specialised agents research, design, build, review and operate real production software, this site included. It is the test bench for everything else on this list.
Invisible governance
Governance-as-code for agentic delivery: claims registers, decision logs, risk registers, review panels and human approval gates enforced inside the workflow, so compliance is a property of the system rather than a stage after it.
AI FinOps and model routing
Cost-aware agent design: routing work between local and cloud models, matching model capability to task value, and making agent spend as observable as cloud spend.
Human-agent team design
Operating models where agents are team members with tiered autonomy. Gremlins is the agent-first SDLC concept: which decisions agents own, which humans gate, and how the review boundary moves as trust is earned.
Context and dependency graphs
Agents are only as good as the context they can reach. Modelling organisations, codebases and decisions as connected graphs that both humans and agents can navigate.
Writing at The Agentic Tekton
The public side of the programme: essays and patterns on agentic operating models, AI-era architecture and technology leadership, published as they survive contact with practice.
Long-running experiments at Coderturtle
The workshop bench: sample architectures, tools and technical writing going back years, now increasingly built by the factory itself.
This work continues the pattern of the sixth career chapter: hands-on experimentation alongside executive leadership, not instead of it.