AI will not fix governance you have not defined
Model-assisted decisions inherit the accountability structure around them. If decision rights are ambiguous today, automation only makes the ambiguity faster.
Thinking, published
Practical pieces on governance, risk and portfolio decision-making under AI. Written for the people who have to defend the decision afterwards.
01Latest
Governance
02All insights
6 articles
Leading indicators beat status reports every time
A RAG status describes a position already reached. A leading indicator describes one you can still change. The difference is where the value sits.
Read
Portfolio decisions are a data latency problem
Most investment committees are not short of judgement. They are short of current information, and they compensate with caution.
Read
Hybrid delivery needs one data model, not one method
The argument about agile versus waterfall is largely settled in practice. What remains unsettled is whether both produce comparable signal.
Read
What an AI-ready PMO actually looks like
It is not the tooling. It is whether the PMO can answer a question it was not asked to prepare for, in the meeting where it matters.
Read
Measure adoption, not deployment
Licences issued is a procurement metric. Decisions changed is a delivery one. Only one of them survives review in a benefits case.
Read
Stay close
New thinking, occasionally.
We publish when we have something worth the read — not on a content calendar.