Liverpool First, Generative Hype Last

Liverpool City Region has been selected as the first test bed for the UK's Industrialising and Digitalising Construction Challenge - a programme backed through UKRI's R&D Missions Accelerator. The pilot will trial digital manufacturing methods, standardised "kits of parts," and improved digital coordination for public infrastructure and social housing delivery.

It's worth being clear about what this actually is. It's not an AI product launch. But that's almost the point. The digital and industrial foundations that this programme is building are exactly what AI-enabled construction workflows need underneath them. You can't run predictive planning or automated coordination on fragmented, paper-based, site-by-site processes. Liverpool is being asked to demonstrate what joined-up delivery looks like when you standardise the approach, share the data, and build with manufacturing logic baked in.

But delivery is what counts. Pilot regions are useful signals, not guarantees. The construction sector will be watching whether the coordination actually holds across design teams, main contractors, and the supply chain - or whether this becomes another well-funded proof of concept that doesn't travel.

CRAYDL published a piece this week arguing that the construction sector is moving away from generic generative-AI experimentation and toward what they're framing as "predictive execution." The emphasis is on digital twins, field coordination, and real-time project intelligence - tools that reduce clashes, rework, delays, and margin erosion rather than tools that generate text or visuals.

This matches what's happening on the ground. The firms getting real value out of AI right now aren't the ones using it to write specifications. They're the ones feeding live programme data, drawing revisions, and RFI logs into systems that flag problems before they become delays. That's a fundamentally different use case - and a much harder one to build for.

But it's worth being honest about where most firms still are. A lot of businesses are still in the "we've tried ChatGPT" phase. Getting from there to genuine predictive capability requires clean data, consistent processes, and people who understand both the technology and the project. Those things take time and they don't appear just because you've bought a platform.

This isn't construction-specific but it's worth tracking for anyone building or procuring construction tech. OpenAI acquired Astral - the team behind Python tooling like uv and ruff - and folded them into the Codex coding-agent team. Anthropic expanded Claude Code to work through messaging platforms. Cursor released Composer 2, a frontier-class coding model built specifically for software engineering workflows.

The pattern is consistent: AI labs are moving beyond "model API" and toward persistent developer tools and ambient agent access. That has direct implications for the software your supply chain and project teams rely on. The construction tech platforms being built or updated right now are being built in an environment where the underlying coding tools are becoming dramatically more capable.

For construction businesses, the practical read is simpler: software that would have taken months to build or customise will take weeks. Custom integrations, site reporting tools, document parsing - the cost and time of building these is falling fast.

PoweringAI launched this week as a pan-European platform with a specific focus: converting powered brownfield industrial and port sites into "energy park plus data centre" developments. The pitch is that legacy industrial land with existing power infrastructure is exactly what AI-ready facilities need. Rather than building on greenfield, the model is adaptive reuse - repurposing the industrial fabric of post-industrial regions into the infrastructure that runs the next generation of AI systems.

This is genuinely interesting for UK construction. Post-industrial sites in the Midlands, the North, and South Wales have the power connections, the land area, and in some cases the structural bones to support this kind of development. The question is whether planning, grid connection timelines, and specialist contractor capacity can keep up with demand.

Construction is being asked to build the buildings that run the AI. That's a strange loop - but it's a real market, and it's growing.

One of today's email digests raised a point worth noting directly. An article in Inc. made the case that AI shouldn't be used to write thought leadership - that the unique perspectives, lived experience, and genuine foresight that define a thought leader can't be replicated by a system that reorganises existing information.

It's a fair point and a useful corrective. AI can assist with research, structuring, and drafting. But if the voice, the opinion, and the practical judgement aren't yours, the content will feel hollow to anyone who knows the subject. In construction especially - where credibility is built on track record and direct experience - that inauthenticity will be obvious.

Use AI as a tool. Don't let it replace the thinking.