Research, delivery lessons, technical perspectives and practical observations from complex programs across Technology, Data & AI, Finance and Engineering.
Data & AI · Point of View
Moving from a successful prototype to a production capability requires more than model performance. Ownership, controls, integration, economics and operating responsibility become the real program.
Finance · Point of View
A new planning platform is not a result. Cycle time, control evidence, forecast quality, and a finance operating model are.
Cross-practice · Point of View
Governance fails when it is a monthly theatre. It works when decision rights, RAID, and quality gates are part of the weekly operating rhythm.
Technology · Lesson Learned
Debt is delayed change, untestable releases, and a risk register that finance cannot see. Treating it as a developer aesthetic is how programs lose a year.
From the work
Lessons, field notes, and practical constraints that show up on programs—without turning every engagement into a case study.
Technology
The chosen pattern often survives. The reason it was chosen, the alternatives rejected, and the residual risk do not. The next team then re-litigates the same decision under pressure.
Read lessonCross-practice
Risks, assumptions, issues and dependencies only govern a program when each item has an owner, a date, and a path to escalate. A list is not a control.
Read field noteData & AI
The operating model for an AI capability has to be designed with the use case. Waiting for promotion is how nobody is accountable for the decision the model is meant to change.
Read lessonFinance · Project Note
A multi-system finance change does not fail first on calculation. It fails when nobody owns reconciliation, exceptions, source of truth, or the cadence of the pack.
What mattered
Deeper perspectives
Points of view and technical briefs for readers who need the argument, not only the observation.
Data & AI · Point of View
Pipelines without product ownership recreate the warehouse problem with newer tools. Data products need contracts, stewards, and service levels.
Technology · Technical Brief
A recorded decision without an expiry becomes dogma. Constraints change. The ADR should say when the choice must be reviewed—not only why it was accepted.
Recent insights
Every published piece in one place. Filter by practice to find what is relevant to your program.
Data & AIPoint of View
Moving from a successful prototype to a production capability requires more than model performance. Ownership, controls, integration, economics and operating responsibility become the real program.
FinancePoint of View
A new planning platform is not a result. Cycle time, control evidence, forecast quality, and a finance operating model are.
Cross-practicePoint of View
Governance fails when it is a monthly theatre. It works when decision rights, RAID, and quality gates are part of the weekly operating rhythm.
TechnologyLesson Learned
Debt is delayed change, untestable releases, and a risk register that finance cannot see. Treating it as a developer aesthetic is how programs lose a year.
TechnologyLesson Learned
The chosen pattern often survives. The reason it was chosen, the alternatives rejected, and the residual risk do not. The next team then re-litigates the same decision under pressure.
Cross-practiceField Note
Risks, assumptions, issues and dependencies only govern a program when each item has an owner, a date, and a path to escalate. A list is not a control.
Data & AILesson Learned
The operating model for an AI capability has to be designed with the use case. Waiting for promotion is how nobody is accountable for the decision the model is meant to change.
FinanceProject Note
A multi-system finance change does not fail first on calculation. It fails when nobody owns reconciliation, exceptions, source of truth, or the cadence of the pack.
Data & AIPoint of View
Pipelines without product ownership recreate the warehouse problem with newer tools. Data products need contracts, stewards, and service levels.
TechnologyTechnical Brief
A recorded decision without an expiry becomes dogma. Constraints change. The ADR should say when the choice must be reviewed—not only why it was accepted.
Next step
Bring a defined initiative or a constraint that does not yet have an owner.
