Industry
AI in construction
AI in construction: document control, estimating support, project reporting, safety observation triage and risk signals — with data constraints, ROI drivers and documented examples.
Miguel Torres, Founder, Merjora · Updated 3 September 2026
In short
Construction's most reliable AI opportunities are administrative rather than physical: drawing and document control, RFI and submittal handling, estimating support from historical cost data, safety observation triage, and automated project reporting from existing site data. Predictive scheduling and computer-vision progress tracking are real but depend on data maturity that most contractors are still building.
Context
Construction margins are thin and the administrative load per project is heavy — documents, RFIs, variations, compliance records and reporting. That administrative layer is where AI earns its place first.
Data is fragmented across project systems, subcontractors and site tooling, so opportunity selection should favour workflows whose inputs already sit in one system.
Where the opportunities are
Document and drawing control
Classifying, versioning and routing project documents, and answering 'which revision governs this detail' from the controlled set.
Complexity: Moderate
RFI and submittal handling
Drafting responses grounded in the specification and prior RFIs, with the responsible engineer approving. Cycle time here directly affects programme.
Complexity: Moderate
Estimating and tender support
Extracting scope items from tender documents and matching them against historical cost data to produce a first-pass build-up for estimator review.
Complexity: High
Safety observation triage
Structuring free-text site observations, clustering recurring hazards and escalating patterns rather than individual reports.
Complexity: Low
Project reporting
Compiling weekly and monthly project reports from existing programme, cost and site data, with commentary drafted for the project manager.
Complexity: Low
Contract and variation review
Surfacing obligations, deadlines and change entitlements from contract documents for commercial team review.
Complexity: Moderate
Constraints that shape delivery
- Data fragmentation across contractors, subcontractors and site systems
- Document quality varies enormously — scans, markups, mixed revisions
- Safety-related outputs must never be the sole basis for a site decision
- Commercial sensitivity around tender and cost data limits what can leave the organisation
- Site connectivity constrains anything requiring real-time inference
ROI considerations
- Value concentrates in engineer, estimator and PM hours released — high loaded cost per hour
- RFI cycle time has programme value; quantify it with your own delay cost, not a generic figure
- Estimating support pays back through win rate and accuracy, both of which need a control comparison
- Reporting automation is easy value but small; treat it as a supporting line
Documented examples
Publicly documented industry examples. These companies are not Merjora clients.
- BAM Ireland →
Publicly documented digital and AI-supported construction delivery.
Frequently asked questions
- What are realistic AI use cases in construction today?
- Document control, RFI drafting, safety observation triage and project reporting. They rely on text you already hold and do not require new site instrumentation.
- Can AI improve construction estimating?
- It can accelerate scope extraction and first-pass build-ups where historical cost data is well structured. It does not replace estimator judgement on risk, buildability or market conditions.
Where would AI pay off in your construction business?
Merjora maps your workflows, scores each opportunity and quantifies the value range — with the constraints named.
Discover your AI opportunitiesRelated reading
Editorial standard. Merjora publishes analysis, frameworks and publicly documented examples. We do not publish invented statistics, unattributed benchmarks or unverified customer stories.