How to Write an AI Policy for Your Construction Company

A construction company AI policy is a short set of rules for how employees use AI tools like ChatGPT, Claude and Copilot.

It should list approved tools, say what data can never go in, require a person to check anything tied to cost, safety or pay, and name who owns the policy. One page is enough.

Your team is likely using AI already, whether there's a policy or not.

In RSM's 2025 survey, 95% of the 80 construction firms surveyed said they use generative AI. And 29% named data privacy and security as a top challenge.

Why contractors need their own version

Generic AI policies are written for office workers. Construction has extra risks:

• Bids and pricing that competitors would love to see
• Owner and sub contracts with confidentiality terms
• Drawings and specs that belong to the architect or owner
• Employee data like pay rates, SSNs and certified payroll records
• Safety content that crews act on in the field

What to include: 7 sections

1. Purpose. One sentence on why the policy exists.
2. Approved tools. Which AI tools are OK, and which accounts to use (company, not personal).
3. Never share. The data that never goes into an AI tool.
4. Always check. The work a person must review before it's used.
5. Disclosure. When to tell an owner, client or partner that AI helped.
6. Owner. Who answers questions and approves new tools.
7. Review date. When the policy gets updated. AI changes fast, so every six months is a good start.

One-page AI policy template for contractors

Copy this and fill in the brackets to kickstart creating your company's AI policy.

Tips to roll it out to your crews

1. Keep it to one page.

2. Share it at your next all-hands or safety meeting.

3. Show three good examples and three bad ones.

4. Add it to onboarding for new hires.

5. Revisit it every six months.

Pick AI tools that make the policy easy to follow

Some AI tools make these rules easier to follow. When you evaluate one, ask: Does it use our existing permissions? Does it log what it does? Does a person confirm changes?

That's how we built Vyki, our AI payroll assistant in hh2 Time Tracking. She works inside your hh2 permissions. Every action is logged under the user's identity. She proposes changes and waits for confirmation. And she won't give tax advice, classify workers or rule on whether a pay practice is legal.

Common questions

What should a construction company AI policy include?

A construction AI policy should include approved tools, data that can never be shared (like SSNs, bid pricing and confidential contracts), work a person must check (quantities, contracts, safety and payroll), disclosure rules, a policy owner and a review date. Keep it to one page so people read it.

Do small contractors need an AI policy?

Yes. Small contractors often have fewer IT controls, so one pasted spreadsheet can expose employee data or bid pricing. A one-page policy takes an hour to write and gives everyone clear rules.

Can employees use personal ChatGPT accounts for work?

Most companies should say no. Personal accounts are outside company control, and you can't see or manage what's shared. Approve company accounts and check their data settings with IT.

How often should an AI policy be updated?

Every six months is a good starting point. AI tools, features and data settings change often. Review the approved tools list, the never-share list and any new use cases your team has found.

Next steps

Copy the template, fill in the brackets, and share it at your next team meeting. When you look at new AI tools, check them against the same rules.