Modern Artificial Intelligence for Contractors and Builders
What the systems sold to contractors as "AI" actually are, and why the difference changes how you check them. This module separates rules, machine learning, computer vision and generative AI, explains how a language model builds an answer, opens up the retrieval, tools and agents inside construction products, and sets out what the published evidence does and does not show.
-
Rules, Machine Learning, Computer Vision and Generative AI on a Project
35 min
-
How a Language Model Builds an Answer
40 min
-
Retrieval, Tools and Agents Inside Construction Software
40 min
-
A Field Guide to Construction AI Tool Categories and the Evidence About Use
45 min
What AI Can and Cannot Reliably Do With Drawings, Specifications and Numbers
Fluent is not the same as right. This module works through the four ways a confident answer goes wrong on a project — an invented specification section, a code answer for an edition nobody adopted, a quantity that does not tie to the takeoff, and a clause read out of a contract nobody executed — and builds the habit of doubt that catches all four.
-
Invented Specification Sections, Citations and Clauses
40 min
-
Adopted Codes, Amendments and Why a Confident Code Answer Can Be Wrong
40 min
-
Quantities, Long Specification Books and Prompt Sensitivity
40 min
-
Automation Bias and the Discipline of Doubt
40 min
Project Data, Customer Privacy and What May Be Pasted Into a Tool
Owner drawings, subcontractor pricing, employee records, homeowner payment details, fingerprint time clocks and recorded site walks all sit on the same laptop, and they do not carry the same rules. This module classifies what a contractor holds, shows where a tool puts it, and builds a decision method for whether a given item may go into a given system.
-
What Counts as Confidential on a Construction Job
40 min
-
Where the Data Goes: Consumer Terms, Business Terms and Retention
40 min
-
Size Thresholds, and the Laws That Reach a Two-Person Remodeller
40 min
-
A Decision Method: May This Go Into This System?
45 min
The Licence, the Seal and the Duties That Do Not Move
A tool can draft, extract and compare. It cannot hold a licence, seal a drawing, supervise construction operations or sign. This module maps who is accountable for each construction output, shows five genuinely different state licensing models so you never mistake one state's rule for the national rule, and draws the line a contractor may not cross into design.
-
Who Is Accountable for What on a Construction Job
40 min
-
There Is No National Contractor Licence
40 min
-
The Line a Contractor May Not Cross Into Design
40 min
-
Supervision, Signatures and the Duties AI Cannot Absorb
40 min
AI-Assisted Code, Contract and Specification Research
An assistant is not an authority. This module builds a research workflow that runs from a well-framed question to a verified answer, and shows why a model code binds nothing until a jurisdiction adopts it, why a standard contract form binds only the parties who sign it, and how to record where an answer came from and who checked it.
-
From Question to Verified Answer
40 min
-
Why a Model Code Is Not Law Until a Jurisdiction Adopts It
40 min
-
Standard Contract Forms Bind Only the Parties Who Sign
40 min
-
Specification References and the Record of Who Checked
40 min
Prompt Engineering for Construction Professionals
A construction prompt that can be checked has ten parts, and the most important of them is the source material you supply instead of asking the tool to remember it. This module builds the structure, shows how to decide what may be pasted, and works through prompt patterns for scope, quotes, requests for information, daily reports, toolbox talks and owner emails.
-
The Ten-Part Construction Prompt
40 min
-
Supply the Source, Never Ask for Recall
40 min
-
Deciding What to Paste, and Reading What Comes Back
35 min
-
Prompt Patterns for Six Everyday Construction Tasks
45 min
Drafting Project Documents, Letters and Records With AI
How to get a first draft of a proposal, RFI, change-event narrative, daily report or owner letter without inheriting its first mistakes. Build a verified project fact file, hunt the invented date, quantity and clause reference, check statute-driven documents against the statute, and keep a record of who prepared, who reviewed and what evidence supports each statement.
-
Drafting From a Verified Project Fact File
40 min
-
Hunting Invented Facts, Dates and Clause References
40 min
-
Documents Whose Contents a Statute Fixes
40 min
-
Records a Named Person Certifies
40 min
AI in Estimating, Takeoff and Bidding
Where AI helps in preconstruction — pulling scope out of specification sections, levelling quotes, drafting proposal narrative — and the line it may never cross: no quantity, price or total leaves your office unless a person can open the document it came from. Plus the statutory freight a public bid carries and the rules that attach to a proposal.
-
Pulling Scope Out of Specification Sections
40 min
-
Every Number Ties Back to a Document
40 min
-
The Statutory Freight a Public Bid Carries
40 min
-
The Licence Number and the Claims in a Proposal
35 min
AI in Safety, Records and the Competent Person
Construction safety duties are written to named human beings: the competent person who inspects, the person who trains and signs the certification record, the engineer whose design a protective system needs, the executive who certifies the summary. This module shows what AI may usefully do around those duties — and why it can never perform one of them.
-
The Named Duty-Holders in a Safety Programme
40 min
-
Training a Person Delivers and Signs
40 min
-
Inspections on the Face of the Work
40 min
-
Recordable Cases, Patterns and What the Research Measured
40 min
AI for Office Operations, Marketing and Customer Work
The office side of a contracting business is where AI reaches customers first: intake, scheduling, proposals, reviews, advertising, calls and texts. This module ranks those uses by the harm a wrong output would do, and works through the federal consumer rules that already govern reviews, endorsements, advertising claims and automated or artificial-voice calls.
-
Lead Intake, Scheduling and Back-Office Document Work
40 min
-
Reviews, Testimonials and Material Connections
40 min
-
Advertising Claims, AI Capability and the Licence Number
35 min
-
Calls, Texts and AI-Generated Voice
40 min
Building Repeatable Construction AI Workflows
A chat window is not a process. This module turns ad hoc AI use into a staged workflow — inputs, data classification, the AI step, verification, reviewer sign-off and record — designed so that a wrong output cannot become a wrong action on site, with the review step built to resist the pull of over-reliance.
-
From Chat Window to Staged Workflow
40 min
-
Designing So a Wrong Output Is Not a Wrong Action
40 min
-
Automation Bias at the Review Step
40 min
-
Worked Workflows and the Workflow Canvas
40 min
AI Agents and Autonomous Systems in a Contracting Business
An assistant that can act is a different thing from one that can answer. This module explains what an agent is, bands agent uses from read-only summarising up to sending a notice or moving money, builds the control points and stop switches that go with each band, and works through separation of duties in an office where three people do everything.
-
What an AI Agent Actually Is
35 min
-
Risk Levels, Control Points and Stop Switches
40 min
-
Excessive Functionality, Permissions and Autonomy
40 min
-
Inbound Content, Residual Risk and Separation of Duties
40 min
Cybersecurity, Data Governance and Evaluating a Construction AI Vendor
How to read what a construction AI vendor actually promises, decide who owns the project data and the model output, and run identity, access, retention, disposal and exit like a contractor rather than an IT department. Built on voluntary federal frameworks, with every limit of each one stated.
-
Reading a Construction AI Vendor's Security and Privacy Claims
40 min
-
Who Owns the Project Data and the Model Output
35 min
-
Identity, Access and Logging From Office to Trailer to Phone
40 min
-
Retention, Device Disposal, Breach and Leaving a Vendor
40 min
Company AI Policy and Quality Control
How a contracting business writes an AI policy people in the office and the field will actually follow, fits AI into the quality and safety programmes it already runs, and keeps the records that show who checked what. No regulator has told contractors what this policy must say, which is exactly why the company has to decide.
-
Why a Contracting Business Needs a Written AI Policy
35 min
-
Approved Tools, Prohibited Data and Acceptable Use
40 min
-
AI Inside the Quality Programme You Already Have
40 min
-
Records, Incidents and Reviewing the Policy
35 min
Practical Construction AI Lab
Four hands-on labs on invented projects: find the errors in an AI-drafted scope and change-event narrative, catch the invented specification section and the citation to a code edition nobody adopted, review a submittal extraction and design the workflow around it, then evaluate a vendor, build a policy and check the marketing copy it produces.
-
Lab One: Errors in an AI-Drafted Scope and Change-Event Narrative
45 min
-
Lab Two: The Invented Specification Section and the Unadopted Code Edition
45 min
-
Lab Three: Submittal Extraction and Designing the Secure Workflow
45 min
-
Lab Four: The Vendor, the Policy and the Marketing Copy Check
45 min
Capstone: The Construction AI Implementation Plan
Build the artefact this Level has been assembling: a written plan for one low-risk first use of AI in a contracting business, with its data, controls, named owners, vendor evidence, measures, failure indicators and stop authority. It promises no savings, because no construction benchmark in this programme's register supports one.
-
Choosing the First Use and Scoping the Plan
45 min
-
Controls, Vendors, Data Classification and Documentation
45 min
-
The Named Owner and the Reviewer for Every Step
40 min
-
Implementation, Baseline Measures, Failure Indicators and Improvement
45 min