Claude for Construction Engineers

Canadian engineering regulators have already set their expectations for artificial intelligence in professional practice. In 2024, Engineers and Geoscientists BC (EGBC) became the first provincial engineering regulator in Canada to issue a formal Practice Advisory on the use of AI, and Professional Engineers Ontario (PEO) adopted that advisory the following year. This course covers the practical side of that guidance. It shows how to apply Claude, the AI assistant from Anthropic, to the document-heavy tasks that fill a construction engineer's week, including drafting site observation reports and non-conformance reports (NCRs), responding to contractor Requests for Information (RFIs), reviewing specifications, preparing change order and delay documentation, and writing short scripts that automate repetitive calculation and reporting work. Every technique is anchored to the Canadian setting, including the 2020 National Building Code of Canada (NBCC), the relevant Canadian Standards Association (CSA) standards, the Ontario Building Code (OBC), and the standard contract forms published by the Canadian Construction Documents Committee (CCDC).

The course builds outward from the lowest-risk tasks. Early modules explain what a large language model is and where its weaknesses come from, including its tendency to produce confident output that is wrong, a behaviour the EGBC advisory addresses under the heading of cognitive bias. The course then teaches a repeatable prompting method built on four parts, being role, context, task, and format, and shows how to structure prompts with XML tags and worked examples. Later modules apply that method to technical report writing, specification and submittal review, code research, Claude Projects for a persistent firm workspace, Python scripting with mandatory verification against a known result, and change management correspondence. A dedicated module covers professional accountability, the governing principles the EGBC advisory sets out for AI use, and a practical AI governance policy for a Canadian engineering firm. The method stays consistent throughout. Claude prepares a first draft and flags what needs checking, and the engineer verifies every factual claim and code reference against the primary source while keeping full responsibility for any work that carries a professional seal.

Learning Objectives

1. Explain how a large language model produces text, and why the same process yields both useful first drafts and confident errors known as hallucinations.
2. Summarize the professional obligations in the Engineers and Geoscientists BC (EGBC) Practice Advisory on artificial intelligence, including its principles of transparency, explainability, and interpretability.
3. Construct prompts using a four-part structure of role, context, task, and format, supported by XML tags and worked examples.
4. Produce first drafts of core construction documentation, including site observation reports, NCRs, RFI responses, and change management correspondence.
5. Distinguish between using AI to interpret contract and code documents and using it to make compliance determinations, and apply that line when researching specifications, CCDC contracts, and NBCC or OBC provisions.
6. Verify AI-generated clause references, standard citations, and calculation outputs against primary Canadian sources before relying on them.
7. Configure a Claude Project and a firm-level AI governance policy that carry Canadian professional context and review standards into daily practice.
8. Implement a staged ninety-day plan for adopting Claude across documentation, specification research, and scripting tasks.

This course is intended for construction engineers and other construction professionals practising in Canada, and it assumes no programming background beyond the ability to read and test a short script. Throughout, Claude is presented as a drafting and research aid rather than a source of professional advice or compliance determinations. It prepares a first pass and flags what needs review, while the verification of every fact and reference, along with the professional accountability that comes with a P.Eng. seal, remains with the engineer.