Every construction company has a few numbers they would never want a competitor to see. The tightest margin you can actually afford to bid. Preferred pricing or labor rates that let you come in just under the next guy. A sequencing trick that saves two weeks on a schedule. You would never hand those over in a meeting with a rival contractor. But if someone on your team has typed them into ChatGPT, or another AI tool, there is a chance you already have.

Most people don’t think about what happens to their information after they hit enter. But hitting enter has consequences.  The free and standard versions of AI tools like ChatGPT are not just answering your question and forgetting about it. Many of these tools are designed to learn from what people type into them. Your prompts and the data you’ve provided become part of the pool of information the AI draws from when it is generating responses for other users down the road. In other words, the very same tool that just helped you write a bid could later be drawn upon to help your competitors write their bids, and it may be pulling on patterns, strategies, or even pricing logic that came from you.

Nobody stole anything. Nobody did anything malicious. That is exactly what makes this characteristic of large language models quieter, and in some ways more unsettling, than a typical data breach. Your own competitive advantage, the pricing strategy you have spent years refining, the way you structure a winning proposal, the efficiencies that let you underbid your competitors could slowly become part of the general knowledge an AI tool uses to help those same competitors in your industry. The construction industry is particularly exposed here because so much of what makes one contractor beat another on a bid is not public information. It is the internal math. It is knowing exactly how tight your margins can get and still turn a profit, or which sequencing tricks shave two weeks off a schedule. Those are not secrets you would ever share with competitors, but they might end up in a prompt to an AI tool without a second thought, because it feels like a private conversation. It is not.

So, what’s the fix? It is not to swear off AI entirely. These tools are genuinely useful and are not going anywhere. The fix is to start treating AI prompts the same way you would treat a conversation with someone outside your company. Before typing something in, ask some simple questions: would I be comfortable if a competitor read this? Have I taken steps to ensure confidentiality? If the answers to these questions are no, then your sensitive information does not belong in a free AI tool. It also helps to draw a clear line between using AI for general help, like tightening up the language in a proposal or summarizing a long document, and using it to process the specific numbers and strategies that make your bids competitive. The first is low risk. The second is where companies get burned without ever realizing it happened. This is also where it may be worth spending a little more money. Many AI companies, such as OpenAI, Anthropic, and Microsoft, now offer business or enterprise versions of their tools that come with a contractual promise not to use your information to train their models. Those paid versions cost more than the free versions everyone is used to, but for a company that is routinely feeding in pricing models or propriety strategies, that extra cost may be a lot cheaper than losing your edge to a competitor.

AI tools are not going away, and they should not have to. Used well, they can save your team real time. But every business owner should adopt this one simple habit: before anything sensitive gets typed into that chat box, pause and ask who else might end up reading it. In an industry where the margins are thin and every edge matters, that pause could be the difference between staying ahead of your competitors and unknowingly handing them your playbook.

Jessica Zaki is about to begin her final year of law school and recently completed a summer clerkship with Dunn DeSantis Walt & Kendrick LLP.

Dunn DeSantis Walt & Kendrick provides a broad spectrum of legal services to businesses of all sizes, from small, local start-ups and non-profits to large, national companies. DDWK’s real estate development and construction practice includes representing all segments of the development and construction industries on both private and public projects. 

You can find additional information and resources related to helping business owners and their businesses on the DDWK website.