---
"title": "AI policy template for a construction firm"
"description": "A plain-English AI acceptable use policy template for
  construction firms, covering approved tools, project and client data, drawings
  and bids, field use, safety-critical work and who owns the policy."
"section": "Playbooks"
"sidebarTitle": "AI policy for construction"
"published": "2026-11-09T14:13:33.733Z"
"updated": "2026-11-09T14:13:33.733Z"
---

## What should an AI policy for a construction firm include?

It depends on how far AI has spread in your firm. If people are only using chat assistants for emails and summaries, a one-page policy covering approved tools, what data is off-limits and who checks output is enough. If estimators, project managers or field teams are uploading drawings, bids or contracts, or you are building agents connected to project systems, you need extra sections on client and owner data, bid confidentiality, safety-critical decisions and access controls. Start short, publish it, and update it quarterly.

## Who this is for

Owners, COOs, general counsel, IT leads and operations managers at general contractors, subcontractors, construction managers and design-build firms who want a usable AI policy without hiring a law firm to write one from scratch.

## The short version

- Name the approved tools and ban personal accounts for company work.
- Classify data into simple tiers: fine to use, approved tools only, and never.
- Keep people responsible for anything that affects safety, pricing, contracts or code compliance.
- Make one person the owner of the policy and review it on a schedule.

This page is a practical template, not legal advice. Have your lawyer review the final version, especially anything tied to contracts with owners.

## Why construction needs its own version

A general [AI acceptable use policy](/docs/write-an-ai-acceptable-use-policy) covers most of the basics. Construction adds a few specifics. Bids and pricing are commercially sensitive and sometimes covered by confidentiality terms in the bid documents. Owner contracts can limit how project information is shared. Drawings and specifications may carry the design team's intellectual property. And some decisions, such as structural, safety and code questions, must stay with licensed or qualified people no matter how good the AI answer looks.

## The template

Copy the sections below and fill in the names, roles and dates that apply to your firm. Keep the finished policy to two pages if you can.

**1. Purpose**

"This policy explains how employees may use AI tools for company work. Its goal is to let people work faster while protecting client, owner and company information and keeping qualified people responsible for decisions."

**2. Approved tools**

List the tools by name and the account type, for example the firm's business plan of a chat assistant, the AI features inside your project management or estimating software, and any internal agents. State that personal and free accounts must not be used for company work. Add the date you last checked each vendor's data terms, because these change. As of November 2026, most major vendors offer business plans that do not train on customer data by default, but check the current terms for the plan you buy.

**3. Data tiers**

- **Tier 1, fine to use in approved tools:** public information, general writing, internal how-to questions, published codes and standards text you are licensed to use.
- **Tier 2, approved tools only, with care:** project schedules, RFIs, submittals, meeting notes, internal cost data and subcontractor correspondence.
- **Tier 3, never without written approval:** bid pricing before award, documents an owner contract marks confidential, employee personal data, and anything a contract says may not be shared with third parties.

**4. Rules for everyone**

1. Use only approved tools and your company login.
2. Check the data tier before uploading a document.
3. Upload only what the task needs, not a whole project folder.
4. Treat AI output as a draft. You are responsible for anything you send or sign.
5. Never rely on AI alone for structural, safety, code compliance or legal decisions.
6. Do not use AI to create or edit drawings, stamps or certifications that require a licensed professional.
7. Report mistakes or accidental uploads to the policy owner the same day.

**5. Estimating and bids**

AI may be used to read specifications, list scope items, compare subcontractor quotes and draft RFIs, using approved tools. Final quantities, pricing and bid submissions must be reviewed by a qualified estimator. Bid pricing stays Tier 3 until award. Our guide to [AI for construction estimating and RFIs](/docs/ai-for-construction-estimating-and-rfis) covers the workflow in more detail.

**6. Field use**

Field staff may use approved mobile apps to draft daily reports, summarize meetings and look up internal procedures. Photos of sites may include people, so follow the firm's rules on images. Safety decisions on site stay with the superintendent and safety lead. See [rolling out AI to construction field teams](/docs/roll-out-ai-to-construction-field-teams) for how to introduce this in practice.

**7. Agents and integrations**

Any AI agent connected to email, document control, the project management system or accounting must be approved by the IT lead, given access only to the projects it needs, and must log what it reads and sends. A person must approve anything an agent sends outside the company. The policy owner keeps a list of active agents and who maintains each one.

**8. Training**

Every employee who uses AI for work completes the firm's AI training, with role-specific sessions for estimators, project managers and field leads. The policy is covered in onboarding.

**9. Ownership and review**

A named person, such as the COO or IT lead, owns this policy and reviews it every three months or when a new tool is approved. Questions go to the policy owner.

## How to roll it out

Publish the policy, then walk teams through it in a short session using real examples from your own projects. A policy that nobody has seen applied to an RFI or a bid is easy to ignore. Pair it with role-based training, such as [AI training for construction project managers](/docs/ai-training-for-construction-project-managers), so the rules show up in daily work.

## Common mistakes

- **Copying a generic policy** that never mentions bids, owners or safety.
- **Banning AI outright,** which pushes people to personal accounts you cannot see.
- **No named owner,** so the approved tools list goes stale.
- **No data tiers,** leaving every upload decision to guesswork.

## How Ruby helps

- **AI strategy and consulting:** adapting this template to your contracts, tools and teams.
- **AI training:** teaching the policy through real construction tasks.
- **AI engineering:** building agents with scoped access, logging and approval steps that match the policy.

If you want a policy your teams will actually follow, [book a call with Ruby](https://rubyadvisory.com/book-a-call).

## Related pages

- [How to write an AI acceptable use policy](/docs/write-an-ai-acceptable-use-policy)
- [AI for construction estimating and RFIs](/docs/ai-for-construction-estimating-and-rfis)
- [How to keep client data safe when using AI](/docs/keeping-client-data-safe-with-ai)
- [Can we use AI on NDA documents?](/docs/can-we-use-ai-on-nda-documents)
