How to find and export your Procore data for AI training

01

What is Procore data, and why do AI labs and data vendors buy it?

What a Procore project records

Commercial construction means buildings such as offices, schools, hospitals and warehouses, and one project can involve dozens of companies. Procore gives each project a set of tools, and every tool keeps one kind of record.

An RFI (request for information) is a written question from a contractor to the architect or engineer about something unclear in the plans, kept together with the written answer. A submittal is a product sample, data sheet or shop drawing that a contractor sends to the designer for approval, and Procore keeps the reviewer's decision beside it.

The daily log is the site manager's record of each day, covering who worked, what was done and what delayed the work. Observations note safety or quality problems seen on site, and the punch list lists the small fixes still owed before the building is handed over. Drawings, photos and documents sit alongside all of these.

Why a resolved RFI teaches an AI model something

An AI model learns by studying a very large number of examples and finding the patterns in them, and those examples are called training data. A closed RFI is a complete example of a problem being handled, because it shows the question, who answered, how long it took and what the drawings changed to as a result. Thousands of them show how the same kinds of conflicts get raised and settled across different buildings.

AI labs now also want to train AI agents, software that carries out a job step by step, and an agent learns a job from records of that job being done. A construction agent could study how submittals move from submission to approval, then be scored on how closely its handling matches the real outcome. The full mechanism is in AI training data: why AI companies want your company's data.

How Procore fits into a dataset

Procore is usually one of several apps in a construction company's dataset, next to email, accounting and scheduling records. Companies are paid $100K to $2M or more per dataset, according to micro1, the data vendor we typically refer companies to. The price covers the whole dataset, not each app on its own, and what your data is worth explains what moves it.

02

Where does Procore show how much data you have?

How to count your projects and the items in them

The steps below follow Procore's guide to exporting a list of projects. Procore has replaced the legacy Company Portfolio page with Procore Home, and the steps cover both.

1. Sign in as a user with Admin permission on the company's Portfolio or Home tool.
2. Open Procore Home (or Portfolio) and set the filters so the list includes inactive projects as well as active ones.
3. On the Projects List card, click the ellipsis (the three dots), choose Export and pick CSV. In the legacy Company Portfolio, click Export above the list instead. The file lands in your browser's default download folder.
4. Count the rows in the CSV. That is your project total.
5. Open a handful of representative projects and note how many items sit in tools such as RFIs, Submittals, Daily Log and Observations. Multiply the average by the project count for an estimated total.

Procore says the legacy Company Portfolio page is being retired, so expect only the Procore Home steps to be available over time.

What each permission level can see

What you see depends on your permission level, and the export runs under the same rules.

Why the total is an estimate

The project list tells you how many projects you have, but not how many items each holds, so the item total is a sample-based estimate. If most of your data is documents, drawings and photos rather than forms, size matters more than item counts. None of the Procore help pages we checked describe a storage readout for the Documents tool, so your Procore account manager is the person to ask for a company storage figure.

03

What should you know about your Procore data before you apply?

The record count

The count is your estimated total of project items, such as RFIs, submittals, daily logs and observations, with the number of projects noted beside it. If most of your Procore data is drawings, documents and photos, give the total size in GB of the folders from a Procore Extracts run instead.

The years of history

The years of history come from the oldest project in your project list. Count from the start date of that project, not from the day your company signed up for Procore, because older projects may have been imported.

The number of seats

A seat is one person at your company with a Procore login. Count your own company's users and leave out outside collaborators, such as the architects and subcontractors you invited into a project.

What you don't need to do first

You don't need to export anything to apply, and you don't need to clean or scrub records. The data vendor anonymizes the records before anyone else sees them, and nothing moves until a lawyer-reviewed license is signed. The application, the intro call and each later step are laid out in how to sell your company's data.

04

How do you export your full Procore history?

Route one: Procore Extracts on Windows

Procore Extracts exports a project's items into a single folder tree and is described in Procore's extract guide. The person running it needs Admin-level permission on every tool they want to extract.

1. Install Procore Extracts on a Windows computer, open it and sign in to Procore.
2. Choose the company and then an active project. Projects set to inactive do not appear.
3. The files go to an Extracts folder in your Documents folder unless you press Change Location. Procore recommends a local drive over a shared one.
4. Tick the tools to extract, or Select All. Include Attachments is on by default, and the Drawings tool has an option to take only the current set.
5. Press Extract and leave the computer on and online until the completion message appears. If the connection drops, Procore retries up to three times, then pauses and resumes when the connection returns.
6. Press View in Folder to open the result, or Perform Another Extract to start the next project.

Route two: Project Archive in the browser

Project Archive is Procore's newer tool, called Data Extracts 2.0 while it was in beta, and it removes the need for the desktop app. Procore's guides to creating and downloading an archive cover it. You need at least Read Only access to each tool you export, and users with Admin permission on the project Directory tool can also see archives other people made.

1. Open the project's Project Archive tool and press Create.
2. Name the archive, choose whether to include a table of contents and select the tools to include.
3. For each tool, optionally set filters, grouping and columns, then tick the items to include, or tick the first column heading to take them all.
4. Press Archive. Procore emails you when it finishes.
5. Return to the Project Archive page and press the download icon. The archive arrives as a .zip file, and the download link stays available for 30 days.

Only archives with a Completed or Partially Completed status can be downloaded. Procore lists more than twenty tools that Project Archive can export, including RFIs, submittals, drawings, daily logs, observations, photos and punch lists, in its tool list.

What the exports contain

Most items come out as PDFs, while documents and photos keep their original file types, such as .xls and .jpeg. Extracts builds a folder tree of project name, date and time, tool and item. If a file path had to be shortened for saving, Extracts adds a File Rename Information CSV that maps each original name to the new one.

What neither route does

Neither route exports every project at once, so a company with many projects runs one export per project. Inactive projects have to be set back to active first, or handled with Procore support. For company-wide storage totals, ask your Procore account manager.

Sources: Procore support, Extract Project Data Using Procore Extracts, Export a List of Projects from Procore Home and Project Archive. Steps checked against these pages in November 2026.

05

What stays private when Procore data is licensed?

You choose what is in scope

Because both routes export tool by tool, sensitive financial records can stay out simply by leaving those tools unticked. Commitments (the contracts with subcontractors and suppliers), prime contract values and bid pricing are the usual ones to leave out. The company can also leave out whole projects.

Names are replaced before a lab sees them

Anonymization means replacing every detail that identifies a person, such as a name or an email address, with a stand-in that is the same every time it appears. An RFI thread then still reads in order, so a lab can learn from it, but it no longer says who wrote it. The method, including what is deleted outright and how you review samples, is in how company data is anonymized for AI training.

Ownership and payment

Licensing gives the buyer limited rights to a copy, while the company keeps ownership of the originals. The terms that matter, such as exclusivity, permitted use and deletion, are covered in data licensing. micro1 pays in full within 30 days of accepting the data, usually as a single payment.