Last week, we looked at where AI may actually fit inside a construction company.
Searching project information.
Summarizing documents.
Comparing revisions.
Organizing information.
Extracting requirements.
Preparing drafts.
There are practical places where AI may be able to assist with the work.
But there’s another side of the conversation.
Your employees may already be finding those opportunities on their own.
Someone in the office may be using AI to help draft an email.
A project manager may be summarizing a document.
An estimator may be asking questions about information they’re reviewing.
Someone in the field may be using an AI tool to help organize notes or work through a question.
The company may never have formally decided to adopt AI.
That doesn’t necessarily mean AI isn’t already being used.
Employees Don’t Have to Wait for the Company to Adopt AI
One of the things that makes AI different from many traditional business technologies is how easy it can be for an individual employee to start using it.
They don’t necessarily need the company to implement a new server, deploy an application, or launch a major technology project.
An employee can open an AI tool and start experimenting.
Often, they’re doing it for a reasonable reason.
They’re trying to get their work done.
They may have heard that AI can help write something, summarize something, answer a question, or find information.
So they try it.
If it helps, they may use it again.
That makes employee adoption an important part of the AI conversation.
The question isn’t simply:
Has our company adopted AI?
It may be:
How is AI already being used inside our company?

Start With Visibility
Before deciding what AI should or shouldn’t be allowed to do, it helps to understand what’s already happening.
Taking a deliberate approach to AI starts with understanding how it’s already being used.
Three questions provide a useful starting point.
1. Which AI tools are people using?
ChatGPT?
Microsoft Copilot?
Claude?
Gemini?
AI capabilities built into software the company already uses?
Something else?
Different employees may be experimenting with different tools for different reasons.
Knowing which tools are being used gives the company a much better starting point than assuming no one is using AI because there hasn’t been a formal AI rollout.
2. What are they using AI to do?
This may be even more important than the name of the tool.
Are employees using AI to:
Draft emails?
Summarize documents?
Work through specifications?
Prepare RFIs?
Organize field notes?
Review estimates?
Research unfamiliar topics?
Create reports?
Search for information?
The answer helps separate useful business applications from experimentation and identify where AI may actually be helping the operation.
It may also reveal opportunities worth supporting more deliberately.
3. What information are they giving it?
This is where the conversation starts to become more complicated.
An employee using AI for a generic question is different from an employee providing project information to get an answer.
Construction companies work with information from many sources.
Drawings and specifications may come from owners, architects, or engineers.
Bid information may come from subcontractors and vendors.
Estimates may contain pricing and other information the contractor considers sensitive.
Project information may involve customers, employees, schedules, costs, contracts, or other business information.
Before deciding how AI should be managed, the company needs to understand what kinds of information employees are putting into these tools.
The Problem Isn’t Necessarily That Employees Are Using AI
It would be easy to make employee AI use entirely about risk.
I don’t think that’s the most useful way to look at it.
If employees are finding ways to use AI to reduce repetitive work, locate information, prepare drafts, or organize information, there may be something worth learning from.
Those employees may be identifying legitimate opportunities that the company hasn’t formally evaluated yet.
The issue is that individual experimentation and company-wide adoption are not the same thing.
An employee may know that an AI tool helps with a task without knowing:
How the tool handles company information.
Whether the company has approved the tool.
What information is appropriate to provide.
What information shouldn’t be provided.
What needs to be verified before the result is used.
Those aren’t necessarily questions an individual employee should be expected to answer on their own.
They are business and technology questions the company needs to address.
A Policy Alone Doesn’t Create Visibility
An AI policy can establish expectations for employees.
But before a construction company can decide what those expectations should be, it helps to understand how AI is actually being used.
Which tools?
For what work?
With what information?
Once those questions are understood, the company can make more informed decisions about appropriate tools, standards, training, and controls.
The exact approach will vary by company.
The important point is that managing AI starts with understanding how it’s already being used.
This Is a Technology Governance Issue
AI is new, but the underlying business problem isn’t entirely new.
Technology can reach the office or jobsite before the documentation, standards, processes, training, and governance needed to support it catch up.
Employees find a useful tool.
They begin using it.
Different people develop different approaches.
Information starts moving through technology the company may not have formally evaluated.
Eventually, the business has to decide what should become standard.
AI makes that cycle especially visible because employees can begin experimenting on their own.
The answer doesn’t have to be to block AI.
And it doesn’t have to be to approve every AI tool employees want to use.
The first step is understanding what is actually happening.
Ask Before You Assume
If your construction company hasn’t formally implemented AI, don’t automatically assume that means employees aren’t using it.
Ask.
Which AI tools are our office and field teams using?
What are they using those tools for?
What company, project, customer, bid, or employee information are they providing?
What guidelines have we given them?
The answers may reveal risks that need to be addressed.
But they may also reveal useful applications that deserve a closer look.
That’s why visibility should come before either unrestricted adoption or a blanket prohibition.
Understand how people are using AI.
Understand what information is involved.
Then decide what should be supported, what should be controlled, and what should become part of the company’s standard way of working.
Construction First. Technology Second.
Better Decisions Keep Projects Moving™.
Understand How AI Is Already Being Used
AI may already be part of your construction company’s operation even if no one has formally decided to adopt it.
Understanding which tools employees are using, what they’re using them for, and what information they’re providing creates a better starting point for deciding what should happen next.
A Construction Technology Review™ looks beyond individual applications to understand how people, processes, project information, and technology work together.
Using the Construction Technology Framework™ and its Operational Foundations, we can look at where AI is already appearing in the operation, where it may provide practical value, and where standards, processes, training, or governance may need to catch up.
The objective isn’t to automatically block AI or approve every tool.
It’s to understand what’s happening first so the company can make better decisions about what comes next.
Before deciding how to manage AI, understand how it’s already being used.

