Your WhatsApp groups are project records. Treat them like it.
The most honest account of what happened on site is sitting in a group chat nobody controls. How to turn WhatsApp traffic into structured, defensible records.
Blog
Notes on AI in construction
Practical, no-hype writing on automation, site reporting, data and adoption.
From denial to leadership on AI
Many construction leaders are still avoiding the AI conversation. The first step is not a strategy or a platform. It is one honest conversation.
AI agents, explained for construction directors
Everyone is selling agents in 2026. What an agent actually is, where they genuinely help a construction business, and the guardrails that stop them causing damage.
Extreme ownership: why "the system failed" is not a sentence
Jocko Willink argues there are no bad teams, only bad leaders. Applied to operations: own your processes, and use automation to see enough to own them.
AI has moved from experiment to infrastructure
The signal in construction AI is no longer the demos, it is the operating decisions: dedicated AI teams, structured training and tools embedded in delivery.
The 7 Habits and the machine that makes the machine
Stephen Covey warned against consuming your production capability to hit today's numbers. Most construction businesses do exactly that.
Copilot is installed. Nobody is using it.
Plenty of firms are paying for Copilot licences that sit idle. Why utilisation stays low, and how to turn a shelfware licence into recovered hours.
Atomic habits for organisations: systems beat goals
You do not rise to the level of your goals, you fall to the level of your systems. For a business, automation is environment design at company scale.
The three AI solutions every UK construction business needs
Site records that write themselves, ISO 19650 compliance on autopilot, and a private GPT over your own documents. Zero new platforms.
The comfort crisis runs businesses too
Michael Easter argues avoiding discomfort quietly weakens us. Businesses do the same, clinging to familiar broken processes. Pick your misogi.
Private AI vs public chatbots: where does your project data go?
Your teams are already pasting project information into AI tools. The question is whether it happens under your rules. Public, enterprise and private AI in plain English.
Make your bed: the case for starting automation small
Admiral McRaven's advice is really about small disciplines done perfectly, every day. It is also the best automation strategy we know.
Golden Thread compliance and Gateway 3 readiness
Gateway 3 is where higher-risk building projects stall. Why manual compliance fails at completion, and how continuous document governance fixes it.
What a good AI pilot looks like
Most AI pilots fail before they start: wrong workflow, no baseline, no owner. The six-week structure that proves value or kills the idea cheaply.
Controlled AI adoption, not blanket bans
What RICS, CIOB, RIBA, the ICE, BSI and government guidance actually say about AI in the built environment: managed use with human oversight.
The site diary is the most underrated document in construction
Nobody gets promoted for writing a good site diary, yet it decides extensions of time and settles arguments years later. Why it deserves automation first.
The real cost of recruiting in the UK built environment
A £100,000 salary is not a £100,000 decision. Once NI, pension, recruiter fees, onboarding and ramp-up are counted, year one can top £147,000.
Meeting notes are free money
Construction runs on meetings, and most of what is agreed in them evaporates. Transcription plus AI turns every meeting into minutes, actions and a searchable record.
Disputes are never born big
Construction disputes start as retrieval problems long before they become legal problems. The case for private, connected commercial intelligence.
Using AI on bids without sounding like a robot
Evaluators are learning to spot AI-written tenders. How to use AI properly in bid production: grounded in your library, edited by your people.
Which AI model should your construction team use?
ChatGPT, Claude, Gemini, Perplexity, Grok and private open-source models: what each is actually for, and where the real savings sit.
MCP and the connected construction office
A quiet technical standard is changing what AI can do: instead of answering questions, it can now operate your tools. What MCP means in practice.
From click-based work to zero-click automation
Software does not fail in construction. Workflows do. How zero-click automation makes the systems you already own do the job you bought them for.
What we learned training sceptical construction teams on AI
Construction people are professionally sceptical, and they are right to be. What actually works when introducing AI to site managers, QSs and estimators.
The real cost of doing nothing about admin
A commercial manager on £70k spending 15 hours a week on automatable tasks costs roughly £26,000 a year in wasted capacity. It never shows on a line item.
We don't do software. We do automation.
Construction does not need another platform. It needs to make the ones it already pays for actually work. Why the money leaks between your systems.