You own your data, and Buildalytic never uses your projects to train a model that serves anyone else. That is the answer construction leaders ask for first, and it is the right question to lead with. It is worth explaining why it matters and exactly how to hold any vendor to it.
Your data is your competitive advantage. A subcontractor's unit pricing, a general contractor's bid strategy, the cost structure behind a buyout, the trades on a job and what they charge. These are trade secrets. When an AI vendor is vague about how it uses your information, that is what is on the line, and a controller or CFO is right to slow the deal down until it is clear.
Do you use our data to train your models?
No. Your data runs your workflows and nothing else. It is never pooled into a shared or public model, and it never makes a competitor's system better. This is worth checking carefully, because many general AI vendor agreements reserve broad rights to use customer inputs to improve their products. That default is the opposite of what a contractor should accept, and the protection belongs in the contract, not a policy page.
This holds all the way down the stack. Buildalytic builds on foundation models from providers such as Anthropic and OpenAI, and their commercial terms are explicit that data sent through their APIs is not used to train their models by default. So your payrolls, bids, and cost structure are not training Buildalytic, and they are not training the model providers underneath it either. You do not have to take our word for it. Read the commitments in the providers' own policies.
The providers say it themselves
Who owns the data, and can we get it back?
You do. You keep your systems of record and your source documents, and you can export everything to CSV or JSON, or delete it, at any time. Ownership is not a setting you toggle later. It is the starting point of the relationship.
Where does it live, and who can reach it?
Each organization's data is isolated from every other customer and encrypted in transit and at rest, in secure United States cloud infrastructure. Access is limited by role and single sign-on, so only the right people and systems ever touch it. Isolation is not an add-on tier. It is how the system is built.
Is our setup proprietary to us?
Yes. Each customer gets its own configured system, grounded on its own data and business rules. Nothing is shared across customers, and the agents built around your workflows are trained for your company alone. Your setup, your data, and your logic are yours.
The one question to ask any AI vendor
Will you commit in writing that our project data is never used to train any model, during the contract and after it ends? If the answer is a privacy policy instead of a clause, keep looking.