Friction Files
What repeats can teach us. What changes can move us.
Patterns, observations and arguments about what gets in the way of turning data, analytics and AI into decisions, action and measurable outcomes.
Recognise the friction
Something familiar?
Most organisational friction does not announce itself as a data, analytics or AI problem.
“We have dashboards everywhere.
Decisions are still slow.”
FILE 002Decision-to-Action LatencyRead the file →
02“Our AI pilots work.
They just don't seem to become operations.”
FILE 012Operational changeRead the file →
03“The same KPI means something different
depending on who you ask.”
FILE 013Definitional driftRead the file →
04“Everyone agrees something should happen.
Nobody quite owns what happens next.”
FILE 011Decision ownershipRead the file →
The Files
Browse the Files.
Follow a recurring problem, a line of thinking or simply whatever catches your attention.
New here? Start with FILE 002 — Decision-to-Action Latency.
Foundations define ideas · Field Notes observe recurring patterns · Perspectives explore wider implications
Everyone agreed on the KPI. Until we asked them to define it.
The slide said active customers. Nobody objected. Then someone asked what counted — and five minutes later there were four definitions on the whiteboard. The organisation had naming consistency. It did not yet have shared meaning.
19 August 2026
The pilot worked. The Tuesday morning didn't change.
The demo was convincing, the output made sense, and the pilot was called a success. Then normal work resumed — same spreadsheet, same queue, same report. A pilot proves that something can work. It does not prove that the organisation can work differently because of it.
18 August 2026
Seven people owned the data. Nobody owned the decision.
Ownership had been defined everywhere — the data, the system, the process, the programme. Then someone asked who decides, and the room went quiet. Data ownership and decision ownership are not the same thing.
17 August 2026
The Missing Link Between Data, AI and Real-World Impact
Why Analytics Engineers, AI Engineers and Forward Deployed Engineers must operate as one connected delivery chain — from reliable data to intelligent systems to operational reality.
28 July 2026
Zero Gravity and IPM — the operating model that finally converts data into movement
Most organisations do not suffer from a lack of data. They suffer from an inability to convert what they know into what they do. Information & Process Management is the practical vehicle that makes Zero Gravity operational — and turns AI ambition into measurable return.
8 July 2026
The New Craft: Designing the People Who Design Intelligent Organisations
Intelligent organisations do not emerge from tools. They are built by a small, deliberate craft — and that craft has to be trained.
20 May 2026
The Smarter AI Becomes, the More Organisational Chaos Becomes Visible
AI does not create the mess. It refuses to hide it. Every contradiction surfaces the moment a model is asked to reason across them.
18 May 2026
Data Was the New Oil. Now It Becomes the New Sun.
Oil was extracted, refined, burned. The Sun is different — abundant, ambient, a force that organises everything around it. Enterprise data must do the same.
12 April 2026
Decision-to-Action Latency: The Hidden Tax on Every Enterprise
The gap between knowing and doing is the most expensive — and least measured — line item on the corporate balance sheet.
8 March 2026
Why Dashboards Are Becoming Liabilities
A dashboard answers yesterday's question for tomorrow's decision. In an agentic world, that gap is no longer survivable.
14 February 2026
The Semantic Layer Is the New Operating System
When meaning becomes a shared, queryable substrate, every application above it becomes interchangeable.
22 January 2026
Agentic Orchestration and the End of the Approval Chain
When agents reason from shared context, the approval chain stops being a control mechanism and starts being a bottleneck.
4 December 2025
Trust Engineering: Designing AI Systems Worth Believing In
Trust is not a feature you add at the end. It is a property of the architecture from the first commit.
19 November 2025
Found something uncomfortably familiar?
The point of documenting friction isn’t to become better at describing it. It’s to remove some.