Construction AI Brief
Skanska is training agents on its own experts' decisions, not just rules. Anthropic published a real engineering blueprint for agents that run for hours. And the UKCW debrief landed on one word: trust.
PlanOps automates the planning tasks you’re reading about.
Start free
Today’s context: This brief covers the latest movements in AI tooling, adoption, and signals for construction teams. Read on for what matters and what to focus on.
Skanska has been building a suite of generative-AI tools it calls Expert Sidekicks - purpose-built agents infused with Skanska-specific knowledge. The Safety Sidekick is the visible example: trained on thousands of documents from experienced safety leaders (policies, procedures, EHS manual, OSHA standards) and, crucially, on how those experts reasoned through decisions in the field, not just the rules they applied. Skanska frames the goal explicitly as closing the industry's knowledge gap as experienced people retire.
This is the construction AI trend that matters more than any model launch. Large contractors are building their own domain agents on proprietary corpora rather than waiting for vendors - and the defensibility comes from the data and the captured expertise, not the underlying model. For UK firms, the strategic question follows: what is your equivalent corpus? Decades of project records, RAMS, lessons-learned and senior-engineer judgement are the asset. The model is increasingly a commodity.
Why it matters
Your competitive moat in AI is your proprietary data and your experts' captured judgement. Start treating your document archive and your senior people's reasoning as a training asset, not just a record.
The Construct UK panel at UK Construction Week - "AI or Die: Rewriting the Rules of Construction Sales and Marketing", chaired by Susan Bendall and featuring voices from Peak One Partners, Building Radar, CBAi and Construct Virtual - played to a full-capacity room on the Marketing & Procurement Stage. The sharpest part of the debate was not about capability. It was about trust, data security and the cultural resistance teams hit when introducing new tooling, alongside practical examples of AI saving time, improving targeting and generating leads.
That framing is the most useful takeaway from the whole week. The market has moved past "should we use AI" to "how do we govern it and get people to trust it". For anyone selling or deploying construction AI, the objections you will hit are now overwhelmingly about trust and security, not whether the tool works.
Why it matters
Lead your AI conversations - internal and client-facing - with trust, data residency and change management. The capability case is largely won; the trust case is where deals and rollouts now stall.
Automate your programme admin. Get your evenings back.
Anthropic engineers Ash Prabaker and Andrew Wilson published a workshop and accompanying engineering write-up this week on building agents that run coherently for hours. The headline lessons are unusually practical: use a three-agent structure (planner, generator, evaluator); prefer adversarial evaluators over self-evaluation, because self-grading reliably becomes self-praise; and don't rely on naive context compaction to maintain coherence - use structured handoffs and a queryable session log that defers irreversible "which tokens to keep" decisions. Decompose long builds into tractable chunks with structured artefacts carrying context between sessions.
For construction software and IT teams building agents - submittal review, compliance evidence chains, programme-drift monitoring - this is the reliability engineering that separates a demo from production. The single most actionable point: separate "doing" from "judging". An agent that grades its own work will tell you it succeeded when it silently failed (the exact failure mode in IBM's "Harnesses in AI" talk: an agent panicking at a login page and reporting success anyway).
Why it matters
If you're commissioning agentic workflows, require an independent evaluator in the architecture. Self-evaluation is the single most common reason production agents quietly fail.
Claude Opus 4.7 introduced task budgets (public beta) - a parameter that sets a token ceiling across an entire agentic loop, with the model receiving a running countdown across thinking, tool calls, tool results and final output. Rather than cutting off unpredictably when the context window fills, the model sees the budget depleting and wraps up gracefully before hitting the limit. It is a suggestion the model can see and self-regulate against, distinct from the hard max_tokens per-request cap.
A practical pattern surfacing across this week's tooling coverage: NotebookLM (and grounded-assistant tools like it) answer only from sources you provide - PDFs, drawings, specs, O&M manuals, contract documents - rather than pulling from the open web. That grounding is the critical property for construction document work, because it sharply reduces the hallucination risk that has made teams nervous about AI on contractual or compliance-critical material.
For UK firms wrestling with Building Safety Act evidence, Golden Thread documentation, NEC clause interrogation or O&M handover, this is a low-risk entry point. You can ask an AI to find inconsistencies across a 46-page specification or a RAMS pack without it inventing a clause that isn't there - and every answer is traceable to a source you supplied.
Why it matters
If trust is the blocker (and the UKCW debrief says it is), source-grounded assistants are the lowest-risk way to demonstrate value. Pilot one on a real document set this fortnight and measure time saved against the manual baseline.
50 free Intelligence Units. Set up your first project in under 20 minutes. No credit card needed.
Get 50 free Intelligence UnitsDaily practical AI insight for construction teams. What changed, why it matters, and what to ignore.
50 free Intelligence Units — automate your programme admin
We help construction teams turn AI into useful work, not noise. Understanding what’s changing in AI is the first step. Making it work on-site is the real difference.
A big month for UK construction AI starts this week. Digital Construction Week opens on Wednesday, Anthropic shipped a flagship with native multi-agent workflows on Friday, and the company is now valued at $965bn. A practical Monday-morning take on what's worth your time.
Found this useful? Share it.
This directly addresses the problem that scuppered many early agentic pilots: an agent silently burning through quota on a long-horizon task. For construction teams running document-heavy or multi-step agentic workflows on a finite budget, this is the cost-control guardrail that makes a production rollout defensible to finance.
Why it matters
If pilot economics were your blocker on agentic AI, task budgets remove a major source of unpredictability. Retest the cost case for any workflow you paused over runaway token spend.
Digital Construction Week is next week, professional indemnity insurers are starting to write AI out of their policies, and LinkedIn has begun throttling the reach of AI-cadence posts. A practical, slightly less polished brief — by design.
Claude landed inside Bluebeam this week. Anthropic and Microsoft shipped the controls that let agents run inside your perimeter. The RTPI warned the planning system can't keep up, and some PI insurers started writing AI out of cover. Digital Construction Week is next Wednesday.